refactor(telemetry): ship the redact waterfall without built-in rules

The seam keeps the telemetry/redact scrubbing interface but ships no rules
of its own: the innermost next() passes records through unchanged, and
deployments mount their rules as waterfall listeners. As an SDK we cannot
know which patterns are secrets in a given deployment; a shipped list
invites false confidence while catching only known shapes, and false
positives would corrupt exported bodies. Mechanism stays with the seam,
policy moves to the deployment; both READMEs and the Agent Note state the
raw-export default plainly.

The loader-composition e2e now mounts a deployment-style rule fixture and
pins the same wire behavior: secret absent, placeholder present, canonical
log untouched.
This commit is contained in:
kingwl
2026-07-23 11:58:47 +08:00
parent cf2e184112
commit 70febffe1a
18 changed files with 155 additions and 236 deletions

View File

@@ -2,5 +2,5 @@
# side as of the last confirmed-consistent state. Both languages carry equal authority;
# after editing either side, bring the other along and re-record with:
# pnpm run verify-translation-pairing --write
2026-07-23-session-telemetry-otel-revival.md: 1150d363e9db98a39e1188b86dc41fa291edba41
2026-07-23-session-telemetry-otel-revival.zh.md: 76749a59c38ef2b8c5b3f09960ca500c25d8e1a8
2026-07-23-session-telemetry-otel-revival.md: 28a88218566dcfe34a3b99b682a4c21d22e00ce2
2026-07-23-session-telemetry-otel-revival.zh.md: 79428c5fbb76bdd8f3c2d0b856edd712b6a650d0

View File

@@ -10,10 +10,10 @@ Every deployment that wants harness sessions in an observability stack must hand
## Decision
`packages/telemetry/` revives the two reviewed packages under the SDK stance — the harness provides the capability, the deployment configures where records go, and nothing crosses the seam unredacted:
`packages/telemetry/` revives the two reviewed packages under the SDK stance — the harness provides the capability, the deployment configures where records go and owns what leaves in them:
- **`@deepseek-ai/dsh-session-telemetry`** — the seam. `TelemetryBackend` (`emit`/`flush?`/`shutdown`), the service-registered `Telemetry` form, and `TelemetryCoordinator` owning capture: adoption with cursor read-back, the per-append firehose (project → `structuredClone` → redact → `emit`, zero I/O), the fixed first-chunk-per-(turn, step) projection, the `agent/error` relay, and dispose-time `shutdown` records.
- **The `telemetry/redact` waterfall** — the delta over the branch version. Every record passes it before reaching any backend; the innermost `next()` applies a conservative built-in rule set (credential shapes: API keys, GitHub/Slack tokens, AWS/Google keys, JWTs, PEM blocks, URL userinfo), deployments stack stricter rules as listeners, and a throwing rule withholds the record fail-closed. The pattern list is a security invariant, deliberately not configurable. Redaction applies to the exported copy only; the canonical log is never rewritten.
- **The `telemetry/redact` waterfall** — the delta over the branch version. Every record passes it before reaching any backend; the seam ships NO rules of its own — the innermost `next()` is a pass-through, deployments mount their rules as listeners (stacking by transforming `next()`'s return value), and a throwing rule withholds the record fail-closed. Redaction applies to the exported copy only; the canonical log is never rewritten.
- **`@deepseek-ai/dsh-session-telemetry-otel`** — the reference backend: OTel JS SDK log pipeline (`LoggerProvider``BatchLogRecordProcessor` → OTLP/HTTP exporter), configured verbatim through `exporter`/`processor` passthroughs. `exporter.url` is required and validated at load; unmounted or unconfigured, nothing leaves the process.
The boundary axiom holds: the harness's aspect ends at `emit()`. Batching, retry, queueing, and loss policy are the reporting SDK's, configured through passthroughs — delivery is best-effort (at-most-once across a crash), which the READMEs state plainly.
@@ -22,12 +22,12 @@ The boundary axiom holds: the harness's aspect ends at `emit()`. Batching, retry
**Implement the runtime-telemetry RFC's outbox (durable spool, per-sink cursors, at-least-once, a `readCommitted` persistence-seam method).** Deferred, not rejected: the SDK stance makes delivery semantics the reporting SDK's territory, and the OTel SDK's own batch pipeline is the honest default. The outbox is a pure additive layer (the `emit()` contract does not move); revive it when a deployment states a crash-loss requirement telemetry must satisfy.
**Export without built-in redaction, delegating to receiver-side collector processors.** Rejected — this is what legal declined. Receiver-side redaction ships the secret first and scrubs it second; the seam must scrub before bytes leave the process, and a waterfall makes the redaction point auditable and stackable.
**No in-process redaction point, delegating to receiver-side collector processors.** Rejected — receiver-side redaction ships the secret first and scrubs it second. The waterfall puts an auditable, stackable scrubbing point before bytes leave the process; where the branch version (what PR #222 shipped) had no redaction point at all, every record now passes one.
**A configurable pattern list for the default rules.** Rejected: deployment-varying tunables belong in config, but a security invariant does not — weakening the floor should require code, not YAML. Stricter rules stack as `telemetry/redact` listeners.
**A built-in conservative rule set as the waterfall's innermost `next()`.** Rejected: as an SDK we cannot know which patterns are secrets in a given deployment, a shipped list invites false confidence ("redaction is on") while catching only known shapes, and false positives would corrupt exported bodies for consumers who never asked. The seam owns the mechanism; the deployment owns the policy — the innermost `next()` is a pass-through, and rules mount as listeners.
**Map onto OTel spans (GenAI semantic conventions) instead of logs.** Rejected for this revival: the branch implementation's log mapping is reviewed and shipped-shaped; the span model is lossy for forkable, interruptible sessions and belongs to a future consumer with real span queries to serve.
## Consequences
A deployment adds one `cordis.yml` entry with an OTLP endpoint and gets its session stream in any OTel-compatible stack; removing the entry is the opt-out, with no residual state. Credential-shaped substrings never leave the process even on a rule-free deployment, at the cost of a synchronous per-record scrub on the capture path (string-regex over lossless-JSON bodies — bounded by event size, no I/O). Exported bodies can differ from canonical log bytes wherever the placeholder landed, so receivers must not treat telemetry as a byte-exact replica; the log remains the source of truth. Crash durability is explicitly out of scope until the outbox decision above is revisited.
A deployment adds one `cordis.yml` entry with an OTLP endpoint and gets its session stream in any OTel-compatible stack; removing the entry is the opt-out, with no residual state. A rule-free deployment exports records exactly as captured — including any credentials embedded in file contents or command output — so a deployment crossing a trust boundary must mount `telemetry/redact` listeners, and both READMEs state this plainly. Where rules are mounted, exported bodies can differ from canonical log bytes, so receivers must not treat telemetry as a byte-exact replica; the log remains the source of truth. Crash durability is explicitly out of scope until the outbox decision above is revisited.

View File

@@ -10,10 +10,10 @@ Status: implemented
## Decision
`packages/telemetry/` 以 SDK 立场复活这两个经过评审的包——harness 提供能力,部署方配置上报去向,且任何数据未经脱敏不得跨越 seam
`packages/telemetry/` 以 SDK 立场复活这两个经过评审的包——harness 提供能力,部署方配置上报去向并对导出内容负责
- **`@deepseek-ai/dsh-session-telemetry`** —— seam 本体。`TelemetryBackend``emit`/`flush?`/`shutdown`)、服务注册形态的 `Telemetry`、以及拥有捕获侧的 `TelemetryCoordinator`:带游标回读的收养、逐 append 的 firehose投影 → `structuredClone` → 脱敏 → `emit`,零 I/O、固定的每 (turn, step) 首 chunk 投影、`agent/error` 转发、以及 dispose 时的 `shutdown` 记录。
- **`telemetry/redact` waterfall** —— 相对分支版本的增量。每条记录抵达任何 backend 前必经此处;最内层 `next()` 应用保守的内置规则集凭据形状API key、GitHub/Slack token、AWS/Google key、JWT、PEM 块、URL userinfo部署方以监听器堆叠更严规则,抛异常的规则将该记录 fail-closed 扣下。模式列表是安全不变量,刻意不可配置。脱敏只作用于导出副本canonical log 永不改写。
- **`telemetry/redact` waterfall** —— 相对分支版本的增量。每条记录抵达任何 backend 前必经此处;seam 自身不带任何规则——最内层 `next()` 原样透传,部署方以监听器挂载自己的规则(通过变换 `next()` 的返回值堆叠),抛异常的规则将该记录 fail-closed 扣下。脱敏只作用于导出副本canonical log 永不改写。
- **`@deepseek-ai/dsh-session-telemetry-otel`** —— 参考 backendOTel JS SDK 日志管线(`LoggerProvider``BatchLogRecordProcessor` → OTLP/HTTP exporter`exporter`/`processor` passthrough 原样配置。`exporter.url` 必填且加载时校验;未挂载或未配置时,任何数据都不会离开进程。
边界公理保持不变harness 的职责止于 `emit()`。批处理、重试、排队与丢失策略属于 reporting SDK经 passthrough 配置——投递是尽力而为崩溃时至多一次README 对此如实陈述。
@@ -22,12 +22,12 @@ Status: implemented
**实现 runtime-telemetry RFC 的 outbox落盘 spool、每 sink 游标、at-least-once、persistence seam 的 `readCommitted` 方法)。** 推迟而非否决SDK 立场使投递语义归属 reporting SDKOTel SDK 自身的批处理管线是诚实的默认。outbox 是纯增量层(`emit()` 契约不动);待某个部署提出遥测必须满足的崩溃丢失要求时再复活。
**不带内置脱敏直接导出,交给接收端 collector processor。** 否决——这正是法务否掉的方案。接收端脱敏是先把秘密发出去再擦除seam 必须在字节离开进程前擦除,且 waterfall 使脱敏点可审计、可堆叠
**不设进程内脱敏点,交给接收端 collector processor。** 否决——接收端脱敏是先把秘密发出去再擦除。waterfall 在字节离开进程前提供一个可审计、可堆叠的擦除点分支版本PR #222 交付的形态)完全没有脱敏点,如今每条记录都必经其一
**默认规则的模式列表做成可配置。** 否决:随部署变化的调优项应进 config但安全不变量不应——削弱底线应当需要改代码而非改 YAML。更严格的规则以 `telemetry/redact` 监听器堆叠
**在 waterfall 最内层 `next()` 内置一套保守规则集。** 否决:作为 SDK 我们无法预知某个部署里什么模式算秘密,内置列表只覆盖已知形状却会带来"脱敏已开启"的虚假信心,且误报会替从未要求过的消费者破坏导出 body。seam 拥有机制,部署方拥有策略——最内层 `next()` 原样透传,规则以监听器挂载
**映射到 OTel spanGenAI 语义约定)而非日志。** 本次复活否决分支实现的日志映射已经过评审、形态可交付span 模型对可 fork、可中断的会话有损留给将来真正有 span 查询需求的消费者。
## Consequences
部署方在 `cordis.yml` 加一个带 OTLP endpoint 的条目即可把会话流接入任何 OTel 兼容体系;删除条目即退出,无残留状态。即使部署方未配置任何规则,凭据形状的子串也绝不离开进程,代价是捕获路径上每条记录一次同步擦除(对 lossless-JSON body 做字符串正则——受事件大小约束,无 I/O导出的 body 在占位符落点处可能与 canonical log 字节不同,接收端不得把遥测当作字节精确副本;日志仍是唯一事实源。崩溃持久性在上述 outbox 决定重启前明确不在范围内。
部署方在 `cordis.yml` 加一个带 OTLP endpoint 的条目即可把会话流接入任何 OTel 兼容体系;删除条目即退出,无残留状态。未挂载规则的部署导出的记录与捕获时完全一致——包括文件内容与命令输出中内嵌的任何凭据——因此跨信任边界的部署必须挂载 `telemetry/redact` 监听器,两个 README 对此如实陈述。挂载规则后,导出的 body 可能与 canonical log 字节不同,接收端不得把遥测当作字节精确副本;日志仍是唯一事实源。崩溃持久性在上述 outbox 决定重启前明确不在范围内。

View File

@@ -807,20 +807,21 @@ Source: [`packages/core/system-prompt/src/index.ts:35`](../../packages/core/syst
### `telemetry/redact` — waterfall
Redact one outbound record before it reaches the backend. The innermost `next()` applies the seam's conservative default rule set (credential-shape scrubbing); listeners stack stricter rules by transforming its return value, and returning without `next()` replaces the default — the exported record is then only as clean as the replacing rule. Dispatched synchronously on the capture hot path inside the coordinator's containment: a throwing listener withholds that one record (fail-closed) and never reaches the agent loop. Redaction applies to the exported copy only; the canonical session log is never rewritten.
Redact one outbound record before it reaches the backend — the seam's scrubbing extension point. The seam ships NO rules of its own: the innermost `next()` passes the record through unchanged, and with no listener mounted records reach the backend as captured, so exported data is exactly as clean as the rules a deployment mounts. Listeners stack by transforming `next()`'s return value; returning without `next()` replaces everything beneath. Dispatched synchronously on the capture hot path inside the coordinator's containment: a throwing listener withholds that one record (fail-closed) and never reaches the agent loop. Redaction applies to the exported copy only; the canonical session log is never rewritten.
```ts cordis-catalog
/**
* Redact one outbound record before it reaches the backend. The innermost
* `next()` applies the seam's conservative default rule set
* (credential-shape scrubbing); listeners stack stricter rules by
* transforming its return value, and returning without `next()` replaces
* the default — the exported record is then only as clean as the
* replacing rule. Dispatched synchronously on the capture hot path inside
* the coordinator's containment: a throwing listener withholds that one
* record (fail-closed) and never reaches the agent loop. Redaction
* applies to the exported copy only; the canonical session log is never
* rewritten.
* Redact one outbound record before it reaches the backend — the seam's
* scrubbing extension point. The seam ships NO rules of its own: the
* innermost `next()` passes the record through unchanged, and with no
* listener mounted records reach the backend as captured, so exported
* data is exactly as clean as the rules a deployment mounts. Listeners
* stack by transforming `next()`'s return value; returning without
* `next()` replaces everything beneath. Dispatched synchronously on the
* capture hot path inside the coordinator's containment: a throwing
* listener withholds that one record (fail-closed) and never reaches the
* agent loop. Redaction applies to the exported copy only; the canonical
* session log is never rewritten.
* @param record - the candidate record, already the coordinator's own deep
* copy; listeners return a (possibly new) record and must not mutate it.
* @mode waterfall
@@ -828,7 +829,7 @@ Redact one outbound record before it reaches the backend. The innermost `next()`
'telemetry/redact'(record: TelemetryRecord, next: () => TelemetryRecord): TelemetryRecord
```
Source: [`packages/telemetry/session-telemetry/src/index.ts:39`](../../packages/telemetry/session-telemetry/src/index.ts)
Source: [`packages/telemetry/session-telemetry/src/index.ts:40`](../../packages/telemetry/session-telemetry/src/index.ts)
## `tools/*`

View File

@@ -1554,7 +1554,7 @@ flush?(): void
abstract shutdown(): Promise<void>
```
Source: [`packages/telemetry/session-telemetry/src/index.ts:123`](../../packages/telemetry/session-telemetry/src/index.ts)
Source: [`packages/telemetry/session-telemetry/src/index.ts:124`](../../packages/telemetry/session-telemetry/src/index.ts)
## `ctx.tokenMeter` — `TokenMeterService`

View File

@@ -43,7 +43,7 @@ This matrix shows which packages dispatch each harness-owned event and which pac
| `subagent/start` | `emit` | [`packages/subagent/subagent/src/index.ts:130`](../packages/subagent/subagent/src/index.ts) | [`subagent`](../packages/subagent/subagent) (`events.dispatch`) | [`hooks-claude`](../packages/hooks/hooks-claude), [`subagent`](../packages/subagent/subagent) |
| `system-prompt/assemble` | `waterfall` | [`packages/core/system-prompt/src/index.ts:29`](../packages/core/system-prompt/src/index.ts) | [`system-prompt`](../packages/core/system-prompt) (`waterfall`) | [`agent`](../packages/core/agent), [`system-prompt`](../packages/core/system-prompt) |
| `system-prompt/change` | `emit` | [`packages/core/system-prompt/src/index.ts:35`](../packages/core/system-prompt/src/index.ts) | [`system-prompt`](../packages/core/system-prompt) (`emit`) | - |
| `telemetry/redact` | `waterfall` | [`packages/telemetry/session-telemetry/src/index.ts:39`](../packages/telemetry/session-telemetry/src/index.ts) | [`session-telemetry`](../packages/telemetry/session-telemetry) (`waterfall`) | - |
| `telemetry/redact` | `waterfall` | [`packages/telemetry/session-telemetry/src/index.ts:40`](../packages/telemetry/session-telemetry/src/index.ts) | [`session-telemetry`](../packages/telemetry/session-telemetry) (`waterfall`) | - |
| `tools/change` | `emit` | [`packages/core/tools/src/index.ts:143`](../packages/core/tools/src/index.ts) | [`tools`](../packages/core/tools) (`emit`) | - |
| `tools/execute` | `waterfall` | [`packages/core/tools/src/index.ts:113`](../packages/core/tools/src/index.ts) | [`tools`](../packages/core/tools) (`waterfall`) | [`session-checkpoint-policy`](../packages/session-persistence/session-checkpoint-policy), [`timeout-policy`](../packages/timeout/timeout-policy) |
| `tools/post-execute` | `waterfall` | [`packages/core/tools/src/index.ts:125`](../packages/core/tools/src/index.ts) | [`tools`](../packages/core/tools) (`waterfall`) | [`hooks-claude`](../packages/hooks/hooks-claude), [`hooks-codex`](../packages/hooks/hooks-codex), [`repeat-tool-guard`](../packages/guard/repeat-tool-guard), [`spill-policy`](../packages/spill/spill-policy), [`tool-fs-search`](../packages/fs/tool-fs-search), [`workspace-context`](../packages/context/workspace-context) |

View File

@@ -1,8 +1,13 @@
# Test-only composition: session-telemetry-otel through the real Loader/app
# path, exporting to the mock OTLP collector the driver starts (url via env).
# The redact-rule entry models a deployment mounting its own scrub rule on the
# telemetry/redact waterfall — the seam itself ships no rules.
- id: cli-mock-llm
name: './cli-mock-llm.ts'
- id: telemetry-redact-rule
name: './telemetry-redact-rule.ts'
- id: bash
name: '@deepseek-ai/dsh-bash-local'

View File

@@ -0,0 +1,29 @@
import type { Context } from 'cordis'
/**
* Deployment-style redaction rule for the telemetry e2e: scrubs the fixture
* credential from body strings, exactly as a real deployment would mount its
* own rules on the `telemetry/redact` waterfall.
*/
const SECRET = /sk-e2efixture[0-9]+/g
const PLACEHOLDER = '[E2E-REDACTED]'
function scrub(value: unknown): unknown {
if (typeof value === 'string') return value.replace(SECRET, PLACEHOLDER)
if (Array.isArray(value)) return value.map(scrub)
if (value !== null && typeof value === 'object') {
return Object.fromEntries(Object.entries(value).map(([key, entry]) => [key, scrub(entry)]))
}
return value
}
export const name = 'telemetry-redact-rule'
/** Mount the fixture scrub rule onto the redact waterfall. */
export function apply(ctx: Context): void {
ctx.on('telemetry/redact', (_record, next) => {
const record = next()
return { ...record, body: scrub(record.body) }
})
}

View File

@@ -32,6 +32,7 @@
"headless-agent/tests/fixtures/time-context-driver.ts",
"headless-agent/tests/fixtures/time-context-mock-llm.ts",
"headless-agent/tests/fixtures/telemetry-otel-driver.ts",
"headless-agent/tests/fixtures/telemetry-redact-rule.ts",
"acp-agent/tests/snapshots/lsp-definition/workspace/subject.ts",
"tui-agent/tests/fixtures/tui-scripted-llm.ts",
"acp-agent/tests/fixtures/subagent/subagent-acp/mock-delegating-llm.ts",

View File

@@ -1124,8 +1124,8 @@ export const EVENT_API: readonly EventApiEntry[] = [
name: 'telemetry/redact',
mode: 'waterfall',
signature: '\'telemetry/redact\'(record: TelemetryRecord, next: () => TelemetryRecord): TelemetryRecord',
jsDoc: '/**\n * Redact one outbound record before it reaches the backend. The innermost\n * `next()` applies the seam\'s conservative default rule set\n * (credential-shape scrubbing); listeners stack stricter rules by\n * transforming its return value, and returning without `next()` replaces\n * the default — the exported record is then only as clean as the\n * replacing rule. Dispatched synchronously on the capture hot path inside\n * the coordinator\'s containment: a throwing listener withholds that one\n * record (fail-closed) and never reaches the agent loop. Redaction\n * applies to the exported copy only; the canonical session log is never\n * rewritten.\n * @param record - the candidate record, already the coordinator\'s own deep\n * copy; listeners return a (possibly new) record and must not mutate it.\n * @mode waterfall\n */',
summary: 'Redact one outbound record before it reaches the backend.',
jsDoc: '/**\n * Redact one outbound record before it reaches the backend — the seam\'s\n * scrubbing extension point. The seam ships NO rules of its own: the\n * innermost `next()` passes the record through unchanged, and with no\n * listener mounted records reach the backend as captured, so exported\n * data is exactly as clean as the rules a deployment mounts. Listeners\n * stack by transforming `next()`\'s return value; returning without\n * `next()` replaces everything beneath. Dispatched synchronously on the\n * capture hot path inside the coordinator\'s containment: a throwing\n * listener withholds that one record (fail-closed) and never reaches the\n * agent loop. Redaction applies to the exported copy only; the canonical\n * session log is never rewritten.\n * @param record - the candidate record, already the coordinator\'s own deep\n * copy; listeners return a (possibly new) record and must not mutate it.\n * @mode waterfall\n */',
summary: 'Redact one outbound record before it reaches the backend — the seam\'s scrubbing extension point.',
},
{
name: 'tools/change',

View File

@@ -1,6 +1,6 @@
# telemetry/
Outbound session reporting: the telemetry seam plus its OpenTelemetry backend. The design — the boundary axiom (the harness's aspect ends at `emit()`; delivery is the reporting SDK's), the mandatory `telemetry/redact` waterfall, the fixed chunk projection, the handoff cursor, and the operational-record channel — is pinned in [the revival Agent Note](../../.agents/notes/implemented/feature/2026-07-23-session-telemetry-otel-revival.md).
Outbound session reporting: the telemetry seam plus its OpenTelemetry backend. The design — the boundary axiom (the harness's aspect ends at `emit()`; delivery is the reporting SDK's), the `telemetry/redact` waterfall (deployment-mounted rules; the seam ships none), the fixed chunk projection, the handoff cursor, and the operational-record channel — is pinned in [the revival Agent Note](../../.agents/notes/implemented/feature/2026-07-23-session-telemetry-otel-revival.md).
| Package | Role |
|---|---|

View File

@@ -19,7 +19,7 @@ The OpenTelemetry backend for [the telemetry seam](../session-telemetry/) — th
## What leaves the machine
Records carry the seam's REDACTED copy of `event.data` — user and assistant message content, tool arguments and results (command output, file contents), the full system prompt and tool schemas (`request/header`), todo text, compaction summaries, hook `stderrSummary`, and the session `cwd` (a local path) — after the seam's `telemetry/redact` waterfall has scrubbed credential-shaped substrings (see [the seam README](../session-telemetry/README.md#the-redact-waterfall)). Provider credentials never appear regardless: adapter API keys are constructor parameters, not session events, so they are structurally absent from the log and therefore from telemetry. A deployment with stricter requirements stacks `telemetry/redact` listeners or opts out structurally.
Records carry the complete `event.data` as the seam's `telemetry/redact` waterfall returns it — user and assistant message content, tool arguments and results (command output, file contents), the full system prompt and tool schemas (`request/header`), todo text, compaction summaries, hook `stderrSummary`, and the session `cwd` (a local path). The seam ships no redaction rules: with no `telemetry/redact` listener mounted, that is the raw captured copy, so a deployment exporting beyond a trusted boundary mounts its own rules (see [the seam README](../session-telemetry/README.md#the-redact-waterfall)). Provider credentials never appear regardless: adapter API keys are constructor parameters, not session events, so they are structurally absent from the log and therefore from telemetry.
## Field mapping

View File

@@ -3,8 +3,8 @@
* a subprocess (per testing policy, through the same app/boot path a
* deployment uses), run one mocked-model turn with a real bash round trip,
* and assert against what the mock OTLP collector actually received on the
* wire: ledger mirroring, default redaction, ops markers, and the untouched
* canonical log.
* wire: ledger mirroring, the deployment-mounted redact rule applied to the
* exported copy, ops markers, and the untouched canonical log.
*/
import { readFile, readdir } from 'node:fs/promises'
@@ -12,7 +12,6 @@ import { join } from 'node:path'
import { fileURLToPath } from 'node:url'
import { describe, expect, it } from 'vitest'
import { LOADER_SMOKE_TEST_TIMEOUT_MS, runLoaderSmoke } from '@deepseek-ai/dsh-loader-smoke'
import { REDACTION_PLACEHOLDER } from '@deepseek-ai/dsh-session-telemetry'
const driver = fileURLToPath(new URL(
'../../../../examples/headless-agent/tests/fixtures/telemetry-otel-driver.ts',
@@ -25,6 +24,7 @@ const configPath = fileURLToPath(new URL(
const repoTsconfig = fileURLToPath(new URL('../../../../tsconfig.json', import.meta.url))
const FIXTURE_SECRET = 'sk-e2efixture1234567890'
const FIXTURE_PLACEHOLDER = '[E2E-REDACTED]'
interface OtlpLogRecord {
attributes?: { key: string; value: Record<string, unknown> }[]
@@ -84,15 +84,16 @@ describe('session-telemetry-otel through a real headless cordis.yml', () => {
}
expect(records.some(({ scope }) => scope.endsWith('/ops'))).toBe(true)
// Default redaction on the wire: the fixture credential never leaves the
// process, its surrounding prose does, and the placeholder marks the spot.
// The deployment-mounted rule on the wire: the fixture credential never
// leaves the process, its surrounding prose does, and the placeholder
// marks the spot — the seam itself ships no rules.
const wire = JSON.stringify(captures)
expect(wire).not.toContain(FIXTURE_SECRET)
expect(wire).toContain(REDACTION_PLACEHOLDER)
expect(wire).toContain(FIXTURE_PLACEHOLDER)
expect(wire).toContain('prove telemetry with key')
// The canonical session log is never rewritten.
expect(logContent).toContain(FIXTURE_SECRET)
expect(logContent).not.toContain(REDACTION_PLACEHOLDER)
expect(logContent).not.toContain(FIXTURE_PLACEHOLDER)
}, LOADER_SMOKE_TEST_TIMEOUT_MS)
})

View File

@@ -12,7 +12,7 @@ The coordinator registers, all through the composing fiber's effects: `session/c
## The redact waterfall
Every record passes the `telemetry/redact` waterfall between projection and `emit()`nothing reaches a backend unredacted. The innermost `next()` applies the built-in conservative rule set (`applyDefaultRedaction`: credential shapes — API keys, GitHub/Slack tokens, AWS/Google keys, JWTs, PEM blocks, URL userinfo — replaced with `[REDACTED]` in body strings and string attribute values). Listeners stack stricter rules by transforming `next()`'s return value; returning without `next()` replaces the default rule set, and a throwing listener withholds that one record fail-closed inside the coordinator's containment. The built-in pattern list is a security invariant, deliberately not configurable from cordis.yml. Redaction applies to the exported copy only; the canonical session log is never rewritten.
Every record passes the `telemetry/redact` waterfall between projection and `emit()`the seam's scrubbing extension point. The seam ships NO rules of its own: the innermost `next()` passes the record through unchanged, so with no listener mounted records reach the backend exactly as captured, and exported data is precisely as clean as the rules a deployment mounts. Listeners stack by transforming `next()`'s return value; returning without `next()` replaces everything beneath, and a throwing listener withholds that one record fail-closed inside the coordinator's containment. Redaction applies to the exported copy only; the canonical session log is never rewritten.
## The handoff cursor
@@ -37,4 +37,4 @@ None; this package neither assembles nor sends a provider request.
## Known Limitations and Deferred Work
- **Best-effort delivery** — the cursor marks handed-off, not delivered; a session torn down inside a reload window cannot be re-adopted; whatever sits in a backend queue at crash time is lost. A durable outbox (spool, per-sink cursors, at-least-once) is deferred until a deployment states a crash-loss requirement — see [the revival Agent Note](../../../.agents/notes/implemented/feature/2026-07-23-session-telemetry-otel-revival.md).
- **Redaction is shape-based** — the default rules catch known credential shapes, not every secret; a deployment with stricter needs stacks `telemetry/redact` listeners, and exported data is only as clean as the mounted rules.
- **No built-in redaction rules** — with no `telemetry/redact` listener mounted, records leave the process exactly as captured, including any credentials embedded in file contents or command output; a deployment exporting to a shared collector owns its rule set.

View File

@@ -2,10 +2,11 @@
* Capture coordinator: the seam's upstream half. Subscribes to the session
* firehose plus the one live-bus relay (`agent/error`), applies the fixed
* chunk projection, builds logical records, runs each through the
* `telemetry/redact` waterfall, and hands the redacted copy to the backend —
* synchronously, with every handler self-contained so a failing backend can
* never starve other subscribers (cordis `emit` is stop-on-throw) or touch
* the agent loop. Composed by a backend in its constructor.
* `telemetry/redact` waterfall (deployment-mounted rules; pass-through when
* none), and hands the result to the backend — synchronously, with every
* handler self-contained so a failing backend can never starve other
* subscribers (cordis `emit` is stop-on-throw) or touch the agent loop.
* Composed by a backend in its constructor.
*
* @module @deepseek-ai/dsh-session-telemetry/coordinator
*/
@@ -14,7 +15,6 @@ import type { Context } from 'cordis'
import type { Session, SessionEvent } from '@deepseek-ai/dsh-session'
import type { Agent } from '@deepseek-ai/dsh-agent'
import type { TelemetryBackend, TelemetryRecord, TelemetrySeverity } from './index.ts'
import { applyDefaultRedaction } from './redact.ts'
/**
* The handoff cursor: per session, the highest `seq` handed to a backend.
@@ -145,13 +145,13 @@ export class TelemetryCoordinator {
/**
* Run the `telemetry/redact` waterfall over one record and hand the result
* to the backend. The innermost `next` applies the seam's conservative
* default rules, so an unconfigured deployment still never exports raw
* credential shapes; callers run inside {@link contain}, so a throwing
* to the backend. The innermost `next` passes the record through unchanged
* — the seam ships no rules; exported data is as clean as the listeners a
* deployment mounts. Callers run inside {@link contain}, so a throwing
* rule withholds the record instead of reaching the loop (fail-closed).
*/
private handOff(record: TelemetryRecord): void {
this.backend.emit(this.ctx.waterfall('telemetry/redact', record, () => applyDefaultRedaction(record)))
this.backend.emit(this.ctx.waterfall('telemetry/redact', record, () => record))
}
/** Forward the turn-end boundary to the backend's optional flush hint. */

View File

@@ -22,16 +22,17 @@ declare module 'cordis' {
interface Events {
/**
* Redact one outbound record before it reaches the backend. The innermost
* `next()` applies the seam's conservative default rule set
* (credential-shape scrubbing); listeners stack stricter rules by
* transforming its return value, and returning without `next()` replaces
* the default — the exported record is then only as clean as the
* replacing rule. Dispatched synchronously on the capture hot path inside
* the coordinator's containment: a throwing listener withholds that one
* record (fail-closed) and never reaches the agent loop. Redaction
* applies to the exported copy only; the canonical session log is never
* rewritten.
* Redact one outbound record before it reaches the backend — the seam's
* scrubbing extension point. The seam ships NO rules of its own: the
* innermost `next()` passes the record through unchanged, and with no
* listener mounted records reach the backend as captured, so exported
* data is exactly as clean as the rules a deployment mounts. Listeners
* stack by transforming `next()`'s return value; returning without
* `next()` replaces everything beneath. Dispatched synchronously on the
* capture hot path inside the coordinator's containment: a throwing
* listener withholds that one record (fail-closed) and never reaches the
* agent loop. Redaction applies to the exported copy only; the canonical
* session log is never rewritten.
* @param record - the candidate record, already the coordinator's own deep
* copy; listeners return a (possibly new) record and must not mutate it.
* @mode waterfall
@@ -142,4 +143,3 @@ export abstract class Telemetry extends Service implements TelemetryBackend {
}
export { TelemetryCoordinator } from './coordinator.ts'
export { applyDefaultRedaction, REDACTION_PLACEHOLDER } from './redact.ts'

View File

@@ -1,77 +0,0 @@
/**
* Conservative default redaction for outbound telemetry records.
*
* Session-event bodies carry file contents and command output that may embed
* credentials; nothing may cross the seam to a backend unredacted. This module
* is the innermost rule set of the `telemetry/redact` waterfall — always
* applied unless an outer listener deliberately replaces the whole chain. It
* scrubs credential-SHAPED substrings from every string in the record body,
* leaving structure (keys, nesting, surrounding prose) intact. The pattern
* list is a security invariant, deliberately not configurable; deployments
* add stricter rules by stacking `telemetry/redact` listeners.
*
* @module @deepseek-ai/dsh-session-telemetry/redact
*/
import type { TelemetryRecord } from './index.ts'
/** Replacement text substituted for each detected credential-shaped span. */
export const REDACTION_PLACEHOLDER = '[REDACTED]'
/**
* Well-known credential shapes. A match anywhere inside a body string is
* replaced; low-signal values (package names, versions, git SHAs, plain URLs)
* deliberately stay untouched — they are the observability signal.
*/
const SECRET_PATTERNS: readonly RegExp[] = [
/sk-(?:ant-)?[A-Za-z0-9_-]{10,}/g, // DeepSeek / OpenAI / Anthropic API keys
/gh[pousr]_[A-Za-z0-9]{16,}/g, // GitHub personal/oauth/server/refresh tokens
/github_pat_[A-Za-z0-9_]{20,}/g, // GitHub fine-grained PAT
/xox[baprs]-[A-Za-z0-9-]{10,}/g, // Slack tokens
/AKIA[0-9A-Z]{16}/g, // AWS access key id
/AIza[0-9A-Za-z_-]{35}/g, // Google API key
/eyJ[A-Za-z0-9_-]{8,}\.[A-Za-z0-9_-]{8,}\.[A-Za-z0-9_-]{8,}/g, // JWT
/-----BEGIN [A-Z ]*PRIVATE KEY-----[\s\S]*?-----END [A-Z ]*PRIVATE KEY-----/g, // PEM blocks
/\b(?<scheme>[a-z][a-z0-9+.-]*):\/\/[^/\s:@]+:[^/\s:@]+@/g, // URL userinfo credentials
]
/** Replace every known credential shape inside one string. */
function scrub(text: string): string {
let out = text
for (const pattern of SECRET_PATTERNS) {
out = out.replace(pattern, REDACTION_PLACEHOLDER)
}
return out
}
/**
* Deep-scrub every string inside a lossless-JSON value, preserving structure.
* The record body is the coordinator's own `structuredClone` — mutation-free
* rebuilding keeps the exported copy independent of the canonical log either way.
*/
function scrubValue(value: unknown): unknown {
if (typeof value === 'string') return scrub(value)
if (Array.isArray(value)) return value.map(scrubValue)
if (value !== null && typeof value === 'object') {
const out: Record<string, unknown> = {}
for (const [key, entry] of Object.entries(value)) out[key] = scrubValue(entry)
return out
}
return value
}
/**
* Apply the conservative default rule set to one record — the innermost
* `next` of the `telemetry/redact` waterfall. Attribute VALUES are scrubbed
* alongside the body (identity attributes are seam-built and boring, but
* `session.cwd` is caller-supplied); attribute keys are seam-owned constants.
* @param record - the candidate record; not mutated.
* @returns a redacted copy safe to hand to a backend.
*/
export function applyDefaultRedaction(record: TelemetryRecord): TelemetryRecord {
const attributes: Record<string, string | number> = {}
for (const [key, value] of Object.entries(record.attributes)) {
attributes[key] = typeof value === 'string' ? scrub(value) : value
}
return { ...record, attributes, body: scrubValue(record.body) }
}

View File

@@ -1,91 +1,19 @@
/**
* Default redaction rules and the `telemetry/redact` waterfall contract:
* credential shapes scrubbed from bodies and attribute values, structure
* preserved, canonical log untouched, listener stacking/replacement, and the
* fail-closed containment of a throwing rule.
* The `telemetry/redact` waterfall contract: pass-through when no listener is
* mounted, listener stacking and replacement, ops-record coverage, the
* untouched canonical log, and the fail-closed containment of a throwing rule.
*/
import { describe, expect, it } from 'vitest'
import { Context } from 'cordis'
import SessionStore, { SessionId } from '@deepseek-ai/dsh-session'
import {
applyDefaultRedaction,
REDACTION_PLACEHOLDER,
TelemetryCoordinator,
type TelemetryBackend,
type TelemetryRecord,
} from '../src/index.ts'
const SECRETS = {
deepseek: 'sk-abcdef1234567890abcdef',
anthropic: 'sk-ant-abcdef1234567890',
githubPat: 'ghp_ABCDEFGHIJKLMNOPqrstuv12345678',
finePat: 'github_pat_ABCDEFGHIJKLMNOPQRSTuvwx',
slack: 'xoxb-1234567890-abcdefghij',
aws: 'AKIAIOSFODNN7EXAMPLE',
google: 'AIzaSyA-1234567890abcdefghijklmnopqrstu',
jwt: 'eyJhbGciOiJIUzI1NiJ9.eyJzdWIiOiIxMjM0NTY3ODkwIn0.SflKxwRJSMeKKF2QT4fwpM',
pem: '-----BEGIN RSA PRIVATE KEY-----\nMIIEow\n-----END RSA PRIVATE KEY-----',
urlCreds: 'https://user:hunter2@internal.example.com/repo.git',
} as const
function record(body: unknown, attributes: Record<string, string | number> = {}): TelemetryRecord {
return { channel: 'ledger', time: 1, severity: 'info', attributes, body }
}
describe('applyDefaultRedaction', () => {
it('scrubs every known credential shape while preserving surrounding text', () => {
for (const secret of Object.values(SECRETS)) {
const out = applyDefaultRedaction(record(`before ${secret} after`))
expect(out.body, secret).not.toContain(secret.includes('\n') ? 'MIIEow' : secret)
expect(out.body).toContain('before ')
expect(out.body).toContain(' after')
expect(out.body).toContain(REDACTION_PLACEHOLDER)
}
})
it('scrubs URL userinfo credentials but leaves plain URLs alone', () => {
const out = applyDefaultRedaction(record(`${SECRETS.urlCreds} and https://example.com/path`))
expect(out.body).not.toContain('hunter2')
expect(out.body).toContain('https://example.com/path')
})
it('recurses through arrays and objects, preserving structure and non-strings', () => {
const out = applyDefaultRedaction(record({
list: [`key=${SECRETS.deepseek}`, 7, null, true],
nested: { text: SECRETS.githubPat, count: 3 },
}))
expect(out.body).toEqual({
list: [`key=${REDACTION_PLACEHOLDER}`, 7, null, true],
nested: { text: REDACTION_PLACEHOLDER, count: 3 },
})
})
it('leaves low-signal values untouched', () => {
const clean = {
pkg: '@deepseek-ai/dsh-session-telemetry@0.0.1',
sha: '342a4c3a9d3adf13cf4ad33b9f8d6e79170be5e2',
prose: 'ordinary sentence with kebab-case-identifier',
}
expect(applyDefaultRedaction(record(clean)).body).toEqual(clean)
})
it('scrubs string attribute values and keeps numeric ones', () => {
const out = applyDefaultRedaction(record(null, {
'session.cwd': `/home/${SECRETS.aws}/proj`,
'event.seq': 4,
}))
expect(out.attributes['session.cwd']).toBe(`/home/${REDACTION_PLACEHOLDER}/proj`)
expect(out.attributes['event.seq']).toBe(4)
})
it('never mutates its input', () => {
const input = record({ text: SECRETS.slack }, { 'session.cwd': SECRETS.aws })
applyDefaultRedaction(input)
expect((input.body as { text: string }).text).toBe(SECRETS.slack)
expect(input.attributes['session.cwd']).toBe(SECRETS.aws)
})
})
const FIXTURE_SECRET = 'sk-fixture1234567890'
class CollectingBackend implements TelemetryBackend {
records: TelemetryRecord[] = []
@@ -99,49 +27,80 @@ async function setup() {
const backend = new CollectingBackend()
const ctx = new Context()
await ctx.plugin(SessionStore)
await ctx.plugin({
const fiber = await ctx.plugin({
name: 'fake-telemetry',
inject: ['sessions'],
apply: (inner: Context) => void new TelemetryCoordinator(inner, backend),
})
return { ctx, backend }
return { ctx, backend, fiber }
}
describe('telemetry/redact waterfall', () => {
it('applies the default rules when no listener is registered', async () => {
it('passes records through unchanged when no listener is mounted', async () => {
const { ctx, backend } = await setup()
const session = ctx.sessions.create(SessionId('w'))
session.append('user/message', { content: [{ type: 'text', text: `key ${SECRETS.deepseek}` }], source: { kind: 'user' } }, { surfaceOp: 'append' })
session.append('user/message', { content: [{ type: 'text', text: `key ${FIXTURE_SECRET}` }], source: { kind: 'user' } }, { surfaceOp: 'append' })
const body = backend.records[0]!.body as { content: { text: string }[] }
expect(body.content[0]!.text).toBe(`key ${REDACTION_PLACEHOLDER}`)
expect(body.content[0]!.text).toBe(`key ${FIXTURE_SECRET}`)
})
it('keeps the canonical log unredacted', async () => {
const { ctx } = await setup()
const session = ctx.sessions.create(SessionId('log'))
session.append('user/message', { content: [{ type: 'text', text: SECRETS.githubPat }], source: { kind: 'user' } }, { surfaceOp: 'append' })
const logged = session.events[0]!.data as { content: { text: string }[] }
expect(logged.content[0]!.text).toBe(SECRETS.githubPat)
})
it('lets a listener stack a stricter rule on top of the defaults', async () => {
const { ctx, backend } = await setup()
it('applies a mounted rule to every outbound record, ops records included', async () => {
const { ctx, backend, fiber } = await setup()
ctx.on('telemetry/redact', (_record, next) => {
const defaulted = next()
return { ...defaulted, body: { shapeOnly: true } }
const record = next()
return { ...record, body: { scrubbed: true } }
})
const session = ctx.sessions.create(SessionId('rule'))
session.append('user/message', { content: [{ type: 'text', text: FIXTURE_SECRET }], source: { kind: 'user' } }, { surfaceOp: 'append' })
expect(backend.records[0]!.body).toEqual({ scrubbed: true })
// The dispose-time shutdown ops record passes through the same waterfall.
await fiber.dispose()
const ops = backend.records.filter(record => record.channel === 'ops')
expect(ops).toHaveLength(1)
expect(ops[0]!.body).toEqual({ scrubbed: true })
})
it('keeps the canonical log untouched by a mounted rule', async () => {
const { ctx } = await setup()
ctx.on('telemetry/redact', (_record, next) => ({ ...next(), body: null }))
const session = ctx.sessions.create(SessionId('log'))
session.append('user/message', { content: [{ type: 'text', text: FIXTURE_SECRET }], source: { kind: 'user' } }, { surfaceOp: 'append' })
const logged = session.events[0]!.data as { content: { text: string }[] }
expect(logged.content[0]!.text).toBe(FIXTURE_SECRET)
})
it('stacks listeners outermost-first around next()', async () => {
const { ctx, backend } = await setup()
const order: string[] = []
ctx.on('telemetry/redact', (_record, next) => {
order.push('outer-before')
const record = next()
order.push('outer-after')
return { ...record, attributes: { ...record.attributes, outer: 1 } }
})
ctx.on('telemetry/redact', (_record, next) => {
order.push('inner')
const record = next()
return { ...record, attributes: { ...record.attributes, inner: 1 } }
})
const session = ctx.sessions.create(SessionId('stack'))
session.append('user/message', { content: [{ type: 'text', text: 'anything' }], source: { kind: 'user' } }, { surfaceOp: 'append' })
expect(backend.records[0]!.body).toEqual({ shapeOnly: true })
session.append('user/message', { content: [{ type: 'text', text: 'hi' }], source: { kind: 'user' } }, { surfaceOp: 'append' })
expect(order).toEqual(['outer-before', 'inner', 'outer-after'])
expect(backend.records[0]!.attributes).toMatchObject({ outer: 1, inner: 1 })
})
it('a listener that skips next() replaces the default rules', async () => {
it('a listener that skips next() replaces everything beneath it', async () => {
const { ctx, backend } = await setup()
ctx.on('telemetry/redact', record => record)
const inner = { called: false }
ctx.on('telemetry/redact', () => ({ channel: 'ops', time: 0, severity: 'info', attributes: {}, body: 'replaced' } satisfies TelemetryRecord))
ctx.on('telemetry/redact', (_record, next) => {
inner.called = true
return next()
})
const session = ctx.sessions.create(SessionId('veto'))
session.append('user/message', { content: [{ type: 'text', text: SECRETS.slack }], source: { kind: 'user' } }, { surfaceOp: 'append' })
const body = backend.records[0]!.body as { content: { text: string }[] }
expect(body.content[0]!.text).toBe(SECRETS.slack)
session.append('user/message', { content: [{ type: 'text', text: 'hi' }], source: { kind: 'user' } }, { surfaceOp: 'append' })
expect(backend.records[0]!.body).toBe('replaced')
expect(inner.called).toBe(false)
})
it('a throwing rule withholds the record fail-closed without disturbing the log', async () => {