DeepSeek/deepseek-harness
1
0
Fork 0
mirror of https://github.com/deepseek-ai/deepseek-harness synced 2026-08-15 21:04:50 +00:00
Code Issues Packages Projects Releases Wiki Activity

docs: replace model experience tables with sections

Browse Source
This commit is contained in:
Tianyi Cui
2026-07-13 22:26:33 +08:00
parent 646f3d2d0c
commit d77f797d89
57 changed files with 712 additions and 325 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
packages/bash/bash-local/README.md
View File

@@ -27,9 +27,11 @@ Design surveyed against the bash tools of Claude Code, OpenCode, Codex, and pi;
## Model Experience
| Context surface | What the model sees | Token effect |
|---|---|---|
| Bash tool results, indirectly | Through `dsh-tool-bash`, the conversation model sees the data-dependent stdout and stderr tail inside that consumer's exact result wrappers, exit and timeout markers, background-task state, and a spill-file path when full output is available. This backend adds no prompt or schema itself. An unknown task becomes exactly `Error: unknown bash task "<taskId>"`; a pre-spawn cancellation becomes `Error: aborted before spawn: <reason>`. | Zero tokens until a bash tool runs. Foreground output is bounded per stream by `maxOutputBytes`; background reads return only new output, so polling does not repeat already-delivered text. Results remain in history until compaction. |
### Bash tool results, indirectly
**What the model sees**: Through `dsh-tool-bash`, the conversation model sees the data-dependent stdout and stderr tail inside that consumer's exact result wrappers, exit and timeout markers, background-task state, and a spill-file path when full output is available. This backend adds no prompt or schema itself. An unknown task becomes exactly `Error: unknown bash task "<taskId>"`; a pre-spawn cancellation becomes `Error: aborted before spawn: <reason>`.
**Token effect**: Zero tokens until a bash tool runs. Foreground output is bounded per stream by `maxOutputBytes`; background reads return only new output, so polling does not repeat already-delivered text. Results remain in history until compaction.
## Known Limitations and Deferred Work

22
packages/bash/bash-sandbox/README.md
View File

@@ -34,11 +34,23 @@ The keyless consumer-integration proofs are `tests/bwrap.e2e.ts`, `tests/landloc
## Model Experience
| Context surface | What the model sees | Token effect |
|---|---|---|
| Bash tool schema, indirectly | By advertising a confining `sandboxMode`, this backend makes `dsh-tool-bash` expose `sandbox_permissions` with enum `workspace-write` \| `danger-full-access` and `justification`. The backend adds no prompt prose, and the session's effective mode remains unstated. | Small fixed schema increment on requests where `bash` is visible; mode switches add no context tokens. |
| Bash tool result, indirectly | After ordinary bounded output, a denied call appends exactly `[sandbox: file access denied under <mode> mode]`. When escalation is available it next appends `[sandbox: escalation available — retry this exact command once with sandbox_permissions (the narrowest wider mode that suffices) + justification; the approval prompt asks the user]`. A settled background runner failure instead appends `[sandbox: the sandbox runner itself failed under <mode> mode — the command did not run; this is a sandbox problem, not a command failure]`. | Zero additional tokens on an unremarkable allowed run beyond ordinary output. Denial or failure adds the quoted conditional marker, retained until compaction. |
| Bash tool error, indirectly | If no runner can enforce a confined mode, the foreground call fails with code `SANDBOX_UNAVAILABLE` and the exact message `sandbox mode "<mode>" is requested but no sandbox backend is usable on this host; refusing to run the command unconfined. Install bubblewrap or run a Landlock-enforcing kernel (Linux), ensure sandbox-exec is usable (macOS) — Windows has no confinement backend yet — or switch the consumer to danger-full-access.` An execution-time runner failure appends ` Runner failure: <first stderr line>`. | Conditional error text is visible for that call and retained in history until compaction. |
### Bash tool schema, indirectly
**What the model sees**: By advertising a confining `sandboxMode`, this backend makes `dsh-tool-bash` expose `sandbox_permissions` with enum `workspace-write` | `danger-full-access` and `justification`. The backend adds no prompt prose, and the session's effective mode remains unstated.
**Token effect**: Small fixed schema increment on requests where `bash` is visible; mode switches add no context tokens.
### Bash tool result, indirectly
**What the model sees**: After ordinary bounded output, a denied call appends exactly `[sandbox: file access denied under <mode> mode]`. When escalation is available it next appends `[sandbox: escalation available — retry this exact command once with sandbox_permissions (the narrowest wider mode that suffices) + justification; the approval prompt asks the user]`. A settled background runner failure instead appends `[sandbox: the sandbox runner itself failed under <mode> mode — the command did not run; this is a sandbox problem, not a command failure]`.
**Token effect**: Zero additional tokens on an unremarkable allowed run beyond ordinary output. Denial or failure adds the quoted conditional marker, retained until compaction.
### Bash tool error, indirectly
**What the model sees**: If no runner can enforce a confined mode, the foreground call fails with code `SANDBOX_UNAVAILABLE` and the exact message `sandbox mode "<mode>" is requested but no sandbox backend is usable on this host; refusing to run the command unconfined. Install bubblewrap or run a Landlock-enforcing kernel (Linux), ensure sandbox-exec is usable (macOS) — Windows has no confinement backend yet — or switch the consumer to danger-full-access.` An execution-time runner failure appends ` Runner failure: <first stderr line>`.
**Token effect**: Conditional error text is visible for that call and retained in history until compaction.
## Known Limitations and Deferred Work

36
packages/bash/tool-bash/README.md
View File

@@ -60,13 +60,35 @@ Under a sandboxing executor this plugin makes the session's standing mode overri
## Model Experience
| Context surface | What the model sees | Token effect |
|---|---|---|
| System prompt | Every request in this plugin's registration scope contains exactly `Check the [exit code: N] marker on every bash result; investigate failures before moving on.` A sandboxing executor adds no mode statement or switch notice. Scoped tool restrictions can hide the schemas without removing this independently registered section. | Small fixed input cost per request while the plugin is active, unchanged by sandbox mode or mode switches. |
| Tool schemas | The model sees `bash`, `bash_output`, and `bash_kill`; their full descriptions and JSON schemas are the exact output of the definitions in `src/index.ts`. `sandbox_permissions` and `justification` appear on `bash` only when the mounted executor advertises sandboxing. Agent-scoped tool restrictions can remove the definitions for that agent. | Fixed schema cost on every request where the tools are visible; sandbox support adds the escalation fields and its conditional description paragraph. |
| Foreground result | The renderer emits the data-dependent stdout tail, then optional `[stderr]` and the stderr tail. With no output it emits exactly `(no output)`. Conditional lines are exactly `[output truncated; full output: <path-or-(unavailable)>]`, `[sandbox: file access denied under <mode> mode]`, `[timed out after <timeoutMs>ms]`, `[killed by signal: <signal>]`, and `[exit code: <exitCode>]`; the sandbox escalation and runner-failure lines are quoted in [`dsh-bash-sandbox`](../bash-sandbox/README.md). | Zero result tokens before a call. Output is bounded per stream, while each emitted line remains in history until compaction. |
| Background task context and results | Start returns exactly `started background task <taskId>`. Completion injects exactly `background bash task <taskId> finished <status>. Read its output with bash_output.` Reads return only the data-dependent delta or `(no new output)`, optionally `[some output was dropped from memory; full output: <paths-or-(unavailable)>]`, then exactly one of `[status: running]`, `[status: killed]`, `[status: killed by <signal>]`, or `[status: completed, exit code: <exitCode>]`. Kill returns `killed background task <taskId>` or `task <taskId> had already finished`. | Start and status text is small; deltas are data-dependent. The completion notice and every tool result are retained until compaction, but polling does not repeat already-delivered output. |
| Tool errors | Validation and policy failures are normalized as `Error: <message>`. This package's stable messages are `invalid command: expected a non-empty string`, `invalid description: expected a non-empty string`, `invalid timeoutMs: expected a positive number, got <value>`, `invalid escalation: sandbox_permissions requires a justification`, `invalid escalation: justification is only valid together with sandbox_permissions`, `invalid justification: expected a non-empty sentence`, `invalid task_id: expected a string, got <value>`, `task <taskId> belongs to another session`, `sandbox_permissions is not available in this composition (no sandboxing executor to escalate)`, `sandbox escalation to "<mode>" is not strictly wider than this call's current "<mode>" mode`, the approval-availability/rejection/cancellation variants, and `command aborted`. | Only the failing call adds these retained tokens; a rejected escalation does not add command output because the command does not run. |
### System prompt
**What the model sees**: Every request in this plugin's registration scope contains exactly `Check the [exit code: N] marker on every bash result; investigate failures before moving on.` A sandboxing executor adds no mode statement or switch notice. Scoped tool restrictions can hide the schemas without removing this independently registered section.
**Token effect**: Small fixed input cost per request while the plugin is active, unchanged by sandbox mode or mode switches.
### Tool schemas
**What the model sees**: The model sees `bash`, `bash_output`, and `bash_kill`; their full descriptions and JSON schemas are the exact output of the definitions in `src/index.ts`. `sandbox_permissions` and `justification` appear on `bash` only when the mounted executor advertises sandboxing. Agent-scoped tool restrictions can remove the definitions for that agent.
**Token effect**: Fixed schema cost on every request where the tools are visible; sandbox support adds the escalation fields and its conditional description paragraph.
### Foreground result
**What the model sees**: The renderer emits the data-dependent stdout tail, then optional `[stderr]` and the stderr tail. With no output it emits exactly `(no output)`. Conditional lines are exactly `[output truncated; full output: <path-or-(unavailable)>]`, `[sandbox: file access denied under <mode> mode]`, `[timed out after <timeoutMs>ms]`, `[killed by signal: <signal>]`, and `[exit code: <exitCode>]`; the sandbox escalation and runner-failure lines are quoted in [`dsh-bash-sandbox`](../bash-sandbox/README.md).
**Token effect**: Zero result tokens before a call. Output is bounded per stream, while each emitted line remains in history until compaction.
### Background task context and results
**What the model sees**: Start returns exactly `started background task <taskId>`. Completion injects exactly `background bash task <taskId> finished <status>. Read its output with bash_output.` Reads return only the data-dependent delta or `(no new output)`, optionally `[some output was dropped from memory; full output: <paths-or-(unavailable)>]`, then exactly one of `[status: running]`, `[status: killed]`, `[status: killed by <signal>]`, or `[status: completed, exit code: <exitCode>]`. Kill returns `killed background task <taskId>` or `task <taskId> had already finished`.
**Token effect**: Start and status text is small; deltas are data-dependent. The completion notice and every tool result are retained until compaction, but polling does not repeat already-delivered output.
### Tool errors
**What the model sees**: Validation and policy failures are normalized as `Error: <message>`. This package's stable messages are `invalid command: expected a non-empty string`, `invalid description: expected a non-empty string`, `invalid timeoutMs: expected a positive number, got <value>`, `invalid escalation: sandbox_permissions requires a justification`, `invalid escalation: justification is only valid together with sandbox_permissions`, `invalid justification: expected a non-empty sentence`, `invalid task_id: expected a string, got <value>`, `task <taskId> belongs to another session`, `sandbox_permissions is not available in this composition (no sandboxing executor to escalate)`, `sandbox escalation to "<mode>" is not strictly wider than this call's current "<mode>" mode`, the approval-availability/rejection/cancellation variants, and `command aborted`.
**Token effect**: Only the failing call adds these retained tokens; a rejected escalation does not add command output because the command does not run.
## Known Limitations and Deferred Work