effective(session) = findLast(the session own knob events)?.value ?? the
composition-config default. One log-only event per knob, owned by its
domain (bash/sandbox-mode in dsh-bash, approval/policy in dsh-approval),
each exporting the same three-piece kit: the event declaration, a pure
fold, and THE write path — a switch IS its event; no owner service, no
facts map. Restart immunity and multi-session isolation fall out of the
log replay by construction.
Execution follows the fold on both sides: the bash tool stamps
escalation grant > session override > executor default, and the approval
seam prepends the never-gate that auto-rejects before any interactive
answerer. Visibility is two layers per knob: a per-agent prompt section
states the effective value on every request (logged through
request/header*, so what-the-model-was-told replays from the log), and an
agent/pre-step narrator injects at most one coalesced delta notice with
positional attribution (user switch vs operator/config drift). The ACP
bridge advertises one capability-gated select per composable knob with
currentValue folded per session, validates set_config_option against the
closed vocabularies, and anchors idle switches at the next turn
prompt-submit under the turn-enclosure contract.
The tool gate advertises sandbox_permissions (an enum of exactly the modes
STRICTLY WIDER than the mounted executor default — the schema makes a
non-widening request inexpressible) plus a required justification, exactly
when ctx.bash.sandboxMode reports a confining mode at registration:
composition truth, never a dead lever. An escalating call resolves
ctx.approval BEFORE anything executes with the audit-self-contained reason
"escalate sandbox to <mode>: <justification>"; allowed-once stamps the
granted mode onto that one bash request (the seam-level per-call override),
while rejected / cancelled / unavailable and the no-service / no-agent
paths each fail closed with their own error text and execute nothing. The
description teaches the flow only when the fields exist: retry the SAME
command once after a real denial, never preemptively; a rejected
escalation is final. No new session events: the attempt is an ordinary
tool/call, the decision is the approval audit pair, the outcome an
ordinary tool/result whose facts name the mode it ran under.
dsh-bash grows the per-call policy carrier: BashExecRequest.sandboxMode
(request-optional, spec required-but-nullable — the owner pattern; resolve()
is the one explicit defaulting step) and the BashExecutor.sandboxMode
capability fact (undefined in the base class — composition truth the tool
layer can read). dsh-bash-local carries the field verbatim and confines
nothing.
dsh-bash-sandbox extends LocalBashExecutor and hands ctx.sandbox the exact
argv it is about to spawn. A denial is a RESULT FACT (the command RAN;
result.sandbox.denied is orthogonal to exitCode/signal), classified
conservatively against the wrap own dialect; a RUNNER failure outranks
denial — foreground re-throws the structured SANDBOX_UNAVAILABLE, a settled
background task stamps sandbox.runnerFailed — so a broken sandbox never
reads as a failing command and the command never runs unconfined.
dsh-tool-bash renders the markers and teaches the model not to retry around
a policy denial; escalation and per-session switching are staged follow-ups.
The exact-disposer fix (5fbac8be B1) repaired agents.register but left the
same wrapper (return () => void dispose()) at seven sibling sites:
tools.register, tools.restrict, systemPrompt.section/tools/variable,
agents.setFactory, and subagents.registerProvider. A wrapper makes correct
composite usage unrepresentable — the exact disposer cannot be recovered, so
a generator effect yielding it leaves the inner effect disposing as a
CONCURRENT SIBLING on owner unload, silently reproducing B1's ordering
corruption. The exact disposer serves both usages (composite-nestable AND
fire-and-forget callable); all seven now return it, typed
() => Promise<void> | void, with the convention pinned by a discriminating
test: an async-link composite probe that passes with the exact disposer and
observes the sibling unregistration firing mid-drain with a wrapper.
Re-auditing also surfaced that B1 itself SHIPPED a full-lint failure: it
changed register()'s return type without updating cross-file consumers
(agent.spec.ts dispose() statements, tool-bash's disposer list), which the
staged-scoped pre-commit lint never saw — pnpm run lint was red at HEAD.
Those three sites and this change's own fallout are fixed together: tests
now await disposers (stronger — they observe the full unwind), sync
paths void them, and the two annotation sites carry the honest union type.
agents.register's README line had drifted the same way (B1 updated the
JSDoc, not the README) — all seven README signatures now match; services
catalog regenerated.
Review round 2 (tianyicui inline comments):
- dsh-system-prompt itself registers the harness:identity (-100) and
deployment:persona (0) sections — they must survive a swapped loop
plugin, so they leave dsh-agent-loop; the persona text is the plugin's
own validated 'persona' config. The model/cwd variables STAY on the
loop: runtime facts of the agents it drives.
- AgentOptions.systemPrompt is deleted with all its forwarding plumbing:
the app configs' systemPrompt keys become 'persona' routed through
dsh-agent-core (schema = z.intersect of the owners'), the ACP bridge
and tool-subagent stop carrying persona configuration, and subagent
children now render the deployment persona like every other agent.
- Example personas drop transport/interface trivia (ACP, CLI) — facts
irrelevant to the model.
- Root CONTEXT.md removed (not idiomatic); its persona definition was
wrong under the new ownership anyway.
- Docs, READMEs, the prompt-variables RFC, and generated catalogs
updated; new loop test pins the assemble-waterfall escape valve
(an emptied assembly sends NO system field).
One principle: every fact in the assembled prompt has exactly one owner.
- dsh-system-prompt: merge-extensible AssembleContext on assemble();
a variable(name, provider) registry; {{name}} interpolation in
renderPrompt, strict (unknown/valueless/malformed references throw);
duplicate section and variable names rejected; assembly carries
resolved section text + variables through the assemble waterfall.
- dsh-agent declares AssembleContext.agent; dsh-agent-loop registers
the agent:persona section (order 0 - identity renders before tool
guidance) and the model/cwd variables, and drops its string join:
renderPrompt(assembly) IS the full prompt.
- Tool guidance moves to its owners: descriptions carry per-tool
semantics; sections only cross-call habits (tool:bash exit-code
habit at order 105; read's not-shell nudge). todo/subagent need no
section - their descriptions already carry the contract.
- SubagentProvider.inheritsParentContext (spawn/acp false, fork true);
dsh-tool-subagent derives truthful per-provider wording and resolves
the provider at load (backend must be listed first).
- Example personas shrink to identity + behavior with {{model}} (and
{{cwd}} in the ACP tree); the welcome banner stops enumerating tools.
RFC: docs/rfc/implemented/architecture/2026-07-05-prompt-variables-and-tool-guidance-ownership.md
The audit swept every packages/*/* plugin for the new AGENTS.md
convention (no hardcoded tunables in plugins) and exposes each finding
as a defaulted, validated Config field. Defaults are the previously
hardcoded values throughout, so no deployment or golden changes.
- tool-fs (had NO Config): readLimit, readMaxLineLength, readMaxBytes,
readStreamMinSize. The caps thread through ReadToolCaps/ReadWindow —
read-render already documented that the consumer applies the caps, so
they become explicit per-request fields.
- tool-web: searchMaxResults (WEB_SEARCH_MAX_RESULTS stays as the
schemastery default). Also fixes the stale GREP_LIMIT references in
search.ts and the web-capability-seam RFC (no such constant exists).
- bash-local: graceMs (SIGTERM->SIGKILL escalation grace). The
RunInternals.graceMs test seam is gone: graceMs is now a required
SpawnSpec field filled from config, so tests exercise the real
config path and the defaults live in exactly one place.
- subagent-acp: disposeEofGraceMs / disposeGraceMs. The AcpRunSpec
fields become required for the same one-defaulting-layer reason.
- session-persistence-sqlite: journalMode ('wal' default; the
rollback-journal modes serve filesystems where WAL's shared-memory
files do not work, e.g. network mounts).
- hooks-claude + hooks-codex: stderrSummaryMaxChars for the persisted
hook/result stderr summary. The duplicated summarize() helpers merge
into hook-protocol's summarizeStderr(stderr, maxChars), beside the
HookResultRecord field it feeds, with the bound parameterized the
same way runHook's defaultTimeoutMs already is.
- compact-basic: charsPerToken for the token estimator (default 4, the
English-text heuristic; CJK-heavy deployments need ~1-2 or compaction
fires far too late). Also corrects the BasicCompactService class doc,
which claimed defaults the required-field config never had.
- fs-local: deletes the dead STREAM_MIN_SIZE constant and the dead
FsIoInternals.streamMinSize seam — the read-routing bound lives in
the consumer (tool-fs), where it is now config. This is item 1 of
the proposed prune-write-only-fs-surface RFC, annotated accordingly.
Every new field gets range validation (following the existing
assertPositiveFinite pattern), a README row, and tests covering the
configured behavior, the schema default, and load-time rejection.
ImageBlock had no production producer and every consumer dropped it:
the deepseek serializer skipped it, the pi-ai converter skipped it as
unrepresentable, the ACP bridge neither advertises image prompt
capability nor forwards image blocks, and compact-basic charged a flat
85-token estimate and rendered an [image] placeholder. A block
constructed today would silently vanish from the wire — the vocabulary
advertised a capability no path honors, the silent-data-loss shape the
defensive patterns warn against. The only constructors were tests
pinning the skip/estimate branches.
Remove ImageBlock and its ContentBlockMap entry (its cache?: CacheHint
field leaves with it; CacheHint itself and the other two cache? fields
are out of scope). compact-basic loses its explicit image estimate and
placeholder arms (the merge-extensible default arms absorb the case);
the deepseek serializer, pi-ai converter, and ACP codec already handled
image in their default arms, so only their image-naming comments
change. The codec's inbound rejection of ACP-protocol image prompt
content stays — that guards wire content a client can send regardless
of our vocabulary.
Tests that constructed harness image blocks to pin the removed branches
are dropped (the 85-token estimate pin) or retargeted onto plugin-added
block types / other non-text blocks, which the surviving default arms
own. Docs, the type-equiv pastes, and the content-block vocabulary
RFC's block list and multimodal-home consequence are updated in the
same change; the RFC moves to implemented/ and the index is
regenerated. A real multimodal feature reintroduces image via
declaration merging together with the adapter mapping, ACP
advertisement, and compaction pricing that honor it.
Replace the "bag of optional fields" tool-presentation types
(ToolCallPresentation / ToolResultPresentation / ToolTerminal) with a
card-tagged discriminated union — the standing FIXME(tool-presentation).
A tool declares one render intent per call/result and the ACP bridge
switches on `card`:
ToolCallView = generic | terminal | diff
ToolResultView = generic | terminal
The `diff` card is new: fs write/edit now emit an ACP {type:'diff'}
content block (an editor's inline diff), which the old shapes could not
express. The bridge also relativizes a file card's title against the
session cwd (mirroring claude-agent-acp's toDisplayPath) while keeping
locations/diff paths raw, and derives the no-capability fenced console
fallback from a terminal result's output. read gains the window-in-title
(`Read foo.txt (5 - 8)`) and an always-set location line, matching the
reference adapter field-for-field.
Migrates all three producer families (tool-fs, tool-bash, tool-todo) and
the sole consumer (the ACP bridge) together — the source does not compile
piecewise. Adds snapshot coverage for the terminal _meta path (a new
capability-advertising scenario) and re-records the fs goldens to show the
diff cards. Applied-hunk (result-time, context-line) diffs need a new
result/event shape and are a follow-up.
RFC: docs/rfc/implemented/architecture/2026-07-02-tool-render-intent-union.md
Codex review of the reframe found stale "trusted-plugin surface/boundary"
wording still in review-relevant spots the first pass missed:
- docs/rfc/README.md index title for the RFC.
- packages/bash/bash-local/src/run.ts (childEnv JSDoc + SpawnSpec stdin/env
JSDoc + the spawn stdin comment) and src/index.ts (resolve carry-through
comment); run.ts also pointed at a tool-bash README section name that no
longer exists.
- the two bash-local test descriptors (run.spec.ts / executor.spec.ts).
- the tool-bash guard test's `boundary-*` call ids and one "boundary
assertion" comment (renamed to `no-forward-*`).
All reworded to the scrub-is-the-control framing (or neutral wording). The RFC
FILENAME keeps `-trusted-plugin-surface` as a stable id (many links point at it;
the index title and content are corrected). No code or behavior change.
Address review: the "trusted-plugin surface" framing overstated the security
story. A model driving the `bash` tool already has equivalent power to set env
vars and feed stdin through ordinary shell syntax (`FOO=bar cmd`, heredocs), so
the `env`/`stdin` seam fields grant it no new capability — and they cannot
exfiltrate the harness's ambient credentials, because the credential SCRUB in
dsh-bash-local (which strips *KEY*/*SECRET*/*TOKEN* from process.env before the
child sees it) is the actual control, and it works regardless of these fields
(tool-call args are static JSON, never shell-evaluated).
So drop the "dangerous / trusted-plugin boundary" language across the RFC, the
three bash-package READMEs, the bash/src/types.ts JSDoc, and docs/bash.md (both
the type-equiv blocks — kept 1:1 with source — and the prose). The reality that
remains: the `bash` tool doesn't EXPOSE env/stdin as parameters because they'd
be redundant with shell syntax; the fields exist for in-process plugins (the
hooks bridges) to pass a JSON payload + CLAUDE_* vars cleanly. The guard test is
kept but reframed: it catches a future `...args` spread that would silently
forward model input into the post-scrub env merge, NOT a trust wall. No code or
behavior change.
The hooks subsystem runs external hook commands the Claude Code / Codex way:
JSON payload on stdin, context in CLAUDE_PROJECT_DIR / CLAUDE_PLUGIN_ROOT env.
Reusing the ctx.bash seam for that needs two new inputs — but stdin and arbitrary
env are exactly what dsh-bash-local's credential scrub exists to keep away from
model-driven commands. So this adds them as a TRUSTED-PLUGIN surface:
- BashExecRequest + BashExecSpec gain optional `stdin` and `env`. They are plain
optionals on the resolved spec (not required-but-nullable like `owner`): a
missing one means "none", the safe default, not a security footgun.
- dsh-bash-local threads them through resolve/run/start. `env` merges AFTER the
credential scrub, so a trusted caller's explicit entry wins even on a
credential-shaped name — the scrub guards the harness's OWN ambient creds from
model-driven commands, not a trusted plugin. stdin is always a pipe, closed
immediately (with bytes when supplied, empty otherwise — EOF as before); an
EPIPE from a child that exits without reading is swallowed.
- The model-facing dsh-tool-bash NEVER forwards model input into stdin/env (its
request is command/workdir/timeoutMs/signal/owner only). A regression guard
drives the real tool with adversarial args and asserts the request carries
neither field — proven to go red if the consumer ever forwards them.
Configurable scrub (in an earlier sketch) is dropped as speculative: the explicit
`env` field already gives a trusted caller full control, and no caller needs to
broaden the ambient scrub. Documented in a new architecture RFC, the bash.md
type-equiv blocks, and the three bash READMEs.
Codex review of the trace-event fold found two merge-blockers.
Blocker #1 — format version. Folding usage onto assistant/message and removing
the standalone usage/error events changed the persisted SessionEventMap shape,
which per the AGENTS.md "bump the version and reject — don't migrate" policy
requires a backend to reject any non-current log. Centralize the version in an
exported SESSION_FORMAT_VERSION constant (dsh-session), read by both write sites
(Session constructor default, SessionStore.prepare header) and the coordinator's
load-time assertVersion check. The constant is pinned at 0: while unreleased the
on-disk format is unstable/pre-release, so breaking shape churn is absorbed at v0
(no monotonic bump until the first tagged release) and any non-0 log is rejected
on load — no migration. Update every test/fixture/doc that stamps a
currently-written header to the constant, bump the ACP snapshot fixture + golden
headers to v0, and keep the version-rejection test meaningful by switching its
bad value to a clearly non-current 99. AGENTS.md documents both the monotonic
(SQLite SCHEMA_VERSION) and pinned-0 (session log) pre-release stances.
Blocker #2 — restore the late turn-end warn. failTurn now sets the error reason
only while the turn is still open; once turn/end is appended (a throwing
agent/turn-end listener after closeTurn) the reason can no longer reach the
durable log, so the late throw is logged via ctx.logger.warn instead of
vanishing into a futile post-close assignment. A regression test asserts the
warn fires.
Also guard the normal-step assistant/message append with the same
content-or-usage condition as the max-tokens branch (a content-less, usage-less
step records no trace-only row), with a covering test.
Type-only change (brands are zero-cost casts; no runtime/wire impact). Closes
the two gaps in the "brand ids that cross package boundaries" policy and fixes
the dependency direction so a capability package never pulls in an unrelated one.
- Extract the `Branded<B>` primitive into a new standalone type-only package
`@deepseek-ai/dsh-brand` (packages/util/brand) with no harness-package deps.
dsh-llm keeps its owned CallId but imports Branded from dsh-brand; dsh-session,
dsh-agent, and dsh-bash all import Branded from there. dsh-bash depends on
dsh-brand ALONE — never on dsh-llm or dsh-session (the architectural fix: a
generic execution backend must not couple to the LLM or session vocabulary).
- Mint BashTaskId + OwnerToken in dsh-bash and thread them through BashTask.id,
the get/ownerOf/list/readOutput/kill seam, the bash-local generation site, and
the dsh-tool-bash validate/access surface. OwnerToken is a DISTINCT brand from
SessionId so the seam stays decoupled; dsh-tool-bash is the single boundary
that casts SessionId -> OwnerToken.
- Brand at the SOURCE, not via mid-pipeline casts: agent-loop's Config types
agents[].id as AgentId and resumeSessionId as SessionId, so the brand enters
at the config boundary and the inner create()/resume casts disappear (only the
genuinely-new per-run session-id string is cast).
- Stop brand erosion: propagate CallId/SessionId/AgentId to the registry/store
Map keys and public params/exports (SessionStore, AgentRegistry + factory
options, the ACP session-id surface + ToolPresenter CallId map, the
persistence coordinator, invariants pendingCalls, the pi-ai tool-call maps).
- Docs: document BashTaskId/OwnerToken in bash.md (type-equiv re-pasted), point
the Branded type-equiv at dsh-brand, fix stale param types in the session/
agent/bash READMEs, regenerate the cordis catalog + module graph.
Implements docs/rfc/proposed/architecture/2026-06-20-branded-ids.md
The original prune removed BashExecutor.get()/.list() too, but each is a
one-line accessor over the executor's already-tracked tasks map, and removing
them forced dsh-tool-bash's tests onto a ~35-line onTaskDone completion-tracking
harness just to replace the one-line ctx.bash.get(id) lookup. Per the AGENTS.md
"RFCs are proposals, not golden truth" principle, that disproportionate
migration cost is evidence the methods earn their keep — a test harness IS a
consumer programming against the seam.
Restore get()/list() (seam + LocalBashExecutor impl + the bash tests that used
them, dropping the doneFor/trackCompletions scaffolding). The persistence
has()/delete()/deleteStored removal stands — it had only contract-test callers
and no test-ergonomics cost. The RFC is retitled persistence-only with an
implementation note recording the bash revert.
Second Codex pass caught two proposed RFCs that describe the bash seam as it
WAS (with get/list) and would read stale once this prune lands, plus a latent
test-helper edge:
- docs/rfc/proposed/architecture/2026-06-20-branded-ids.md and
2026-06-20-generic-long-running-tool-runtime.md: drop get/list from the
BashExecutor seam description (surviving: resolve/run/start/ownerOf/
readOutput/kill/onTaskDone). branded-ids will be further updated when it is
implemented; this keeps it accurate in the meantime.
- trackCompletions now records every completion to `done` unconditionally (and
also wakes a parked waiter), so a second doneFor(id) after completion resolves
instead of hanging.
Codex review of the PR2 diff caught doc/comment sites doc-sync does not gate
(core-data-structures prose) and a latent test-helper race:
- docs/core-data-structures/persistence.md + bash.md, sqlite README, and two
source comments (coordinator.ts, jsonl.spec.ts) still listed the removed
has/delete/get/list methods — updated to the surviving four-method
persistence surface and the get/list-free bash seam.
- doneFor(ctx, id) attached its onTaskDone listener lazily, after the task
could already have closed (e.g. `true`), so it could miss the completion and
hang. Replaced with trackCompletions(ctx): one eagerly-installed listener
(mounted in setup() before any task starts) records every completion, and
doneFor resolves immediately for an already-finished task or on completion
otherwise. Race-free, and there is no get-by-id seam left to poll instead.
Two capability seams carried abstract methods no production consumer calls.
A method no consumer programs against is not a seam — it is speculative
surface every implementation must still provide and test.
- SessionPersistence: remove has() and delete(), the coordinator's
has/delete/deleteCore, and the PersistenceBackend.deleteStored hook (with its
jsonl + sqlite + in-spec memory-stub impls). Surviving service surface:
create/append/load/list. Production uses only load() (resume) and list()
(ACP session/list).
- BashExecutor: remove get(id) and list(), the abstract decls and the
LocalBashExecutor impls. The internal tasks map survives (it backs
ownerOf/readOutput/kill); get/list were pure public accessors over it with no
shipping caller and no bash_list tool.
- Migrate tests that reached through ctx.bash.get(id) to the public completion
seam: a doneFor(id) helper over onTaskDone awaits a task by id, and the
HMR-reload ownership test now proves task survival through A's own bash_output
([status: running]) plus ownerOf + B-rejection — a stronger through-the-tool
assertion than the removed lookup peek.
- Update seam READMEs (six -> four service methods, drop the deleteStored hook
and the get/list row) and the two implemented persistence RFCs in place.
Implements docs/rfc/implemented/simplification/2026-06-20-prune-dead-seam-methods.md
Move the 18 flat packages/<name> packages into role-grouped dirs:
core/, llm/, bash/, session-persistence/, ui/, support/. Group dirs are
pure containers; each package keeps its @deepseek-ai/dsh-* name.
Collapse the per-package tsconfig paths maps (base + typecheck) into one
@deepseek-ai/dsh-* wildcard with a candidate per group, and derive the
publint list from the hierarchy. Update all depth-coupled globs/configs
(workspace, tsdown, vitest, eslint, knip, tsconfig includes/refs,
per-package tsconfigs, generators, doc-script scopes, type-equiv manifest)
and the cross-package/script relative imports in tests.
Fix doc-typecheck's workspacePaths() to parse tsconfig JSONC via the
TypeScript API instead of a regex comment-strip, which corrupted the
new wildcard `/*/` path candidates.
WIP: doc cross-links and package/RFC docs still to update.
