Commit Graph
8 Commits
Author SHA1 Message Date
Rahul Selvaraj feac4f52ba fix(hooks): match the full classifier notice, isolate test env
CLASSIFIER_NOTICE matched by prefix, so a pasted notice followed by a
user question was dropped from both summary paths. Anchor it to the
whole message so trailing user text is kept, and add a regression test.

Also pass ECC_SKIP_LLM_SUMMARY=1 in the withTranscript fixture helper so
an inherited ECC_LLM_SUMMARY_INTERVAL=1 cannot trigger claude -p.

Refs #3867

Assisted-by: Claude <noreply@anthropic.com>
2026-11-06 08:47:21 +08:00
Rahul Selvaraj 846a561a7b fix(hooks): drop classifier notices from LLM summary input
The safety-classifier notice filter only covered the mechanical summary.
When generateSessionSummary runs (low context or message interval),
extractConversationText read the raw transcript and passed the notice to
the LLM as a user turn, so it could still reach saved context.

Move the check into one shared isClassifierNotice() in
scripts/lib/llm-summary.js and use it in both userMessageText() and
extractConversationText(). Add LLM-path regression tests, and split the
session-end classifier-notice tests into small functions (<50 lines).

Refs #3867

Assisted-by: Claude <noreply@anthropic.com>
2026-11-05 18:06:51 +08:00
haelyra b37942d3c1 fix(summary): preserve work before the earlier compaction boundary
Separate confirmed model capacity from Stop-summary scheduling. Use the smaller confirmed capacity/native budget, or native-only budget for inferred capacity. Three unit and one actual hook RED cases now pass; helper43, actualStop15 and ten independent controls pass without changing the capacity resolver.
2026-10-29 17:37:52 -07:00
haelyra fc549b6372 fix: preserve Windows rule examples and conservative summary budgets
Address #3809 review regressions for CRLF code fences and inferred context capacity. Reproduced unit and actual Stop-hook failures; verified llm-summary38, Stop14, Copilot37 and independent code/security review. Keep confirmed model/ECC context authority, native budget scheduling and unknown capacity separate.
2026-10-29 17:27:56 -07:00
Tivonsico 1ad235fdfe fix: report explicitly disabled summaries as skipped 2026-10-11 00:36:27 +08:00
Tivonsico 71bf78796f fix: keep session summaries within the lifecycle deadline 2026-10-11 00:15:12 +08:00
haelyra 950caaaae1 fix(hooks): preserve short sessions and quote evolved metadata 2026-08-28 16:08:19 -04:00
c2950121c9 feat(session): LLM-powered session summary via claude -p (#2388)
Replace mechanical text extraction in session-end.js and pre-compact.js
with LLM-generated summaries using `claude -p`. Summaries now capture
design decisions, resolved bugs, changed files, and carry-over context
rather than just truncated user message snippets.

- Add scripts/lib/llm-summary.js: generateSessionSummary, extractConversationText,
  getContextRemainingPct, getContextThreshold, getLLMModel
- Update scripts/hooks/session-end.js: trigger LLM when context < 20% or
  every 50 messages (env-configurable via ECC_LLM_SUMMARY_*)
- Update scripts/hooks/pre-compact.js: generate LLM summary right before
  compaction and write it to the active session .tmp file
- Add tests/lib/llm-summary.test.js: 18 unit tests
- Update tests/hooks/hooks.test.js: 3 integration tests for new behaviour

Recursion guard: sets ECC_SKIP_LLM_SUMMARY=1 in subprocess env so Stop
hooks fired by the claude -p subprocess do not re-enter summarisation.
Requires no ANTHROPIC_API_KEY — reuses Claude Code's own authentication.

Co-authored-by: Hiroshi Tanaka <hiroshi_tanaka@MBAM3.local>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-06-29 15:55:01 -07:00