Coder/deepseek-harness
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: expose agent session log location

Browse Source
This commit is contained in:
Yichen Jiang
2026-07-10 20:52:27 +08:00
parent 42ebbfdf8f
commit eea0a99985
40 changed files with 526 additions and 70 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
packages/hooks/hooks-codex/README.md
View File

@@ -48,6 +48,8 @@ The hooks themselves run in the agent's session workspace: for the agent-scoped
A tool call's payload carries the real `tool_name` (the same value the matcher tests) and Codex's `tool_input: { command }` shape (the `command` arg when present, else `''`). The matcher subject is the tool name (`PreToolUse`/`PostToolUse`) or the session source (`SessionStart`); `UserPromptSubmit`/`Stop` ignore matchers.
Every agent-scoped stdin payload carries `session_id` and `transcript_path`. The bridge resolves the latter through `ctx.sessionPersistence.locate(session.header)` when available and otherwise sends `null`, preserving the Codex `string | null` shape. Lookup does not create or flush the artifact, so a path can be absent before the first turn-end checkpoint or omit the current open turn.
`SessionStart` — the one emit point — runs detached; each run chain is tracked, and disposing the bridge aborts a still-running hook process, then drains the continuation before the dispose resolves (`createDetachedRuns` in `dsh-hook-protocol`).
## Context source