feat(gaia): reliable, cycle-aware cost accounting for plan execution (SPEC-017)#539
Merged
Merged
Conversation
…g hook Add .gaia/scripts/token-rollup.sh: reads the token ledger, dedups each session's re-invocation rows (non-partial, max total, tiebreak latest ended_at), and renders the spec/plan/execute/total cycle breakdown with summed active-span elapsed and the four billing buckets. Always exits 0; degrades to a partial/absent figure with a marker, never fabricates. Add .claude/hooks/token-tally-git-op.sh (PreToolUse Bash) plus shared resolver lib .claude/hooks/lib/gaia-active-plan.sh: record this execution session's ground-truth tally on the orchestrator's per-phase git commit/push, gated on an active plan folder, keyed to the feature. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Contributor
|
code-review-audit skipped: no audit-relevant files changed in the un-audited delta (since the last clean audit); GAIA-Audit commit status stamped on HEAD so the merge gate is satisfied with no local audit run |
Add .claude/hooks/token-rollup-merge.sh (PostToolUse Bash): on gh pr merge, resolve the feature key from the active plan folder (or the ledger's most recent execute row as a labeled fallback) and render the full-cycle spec/plan/execute/total roll-up into the merging session. Register both new hooks in .claude/settings.json (token-tally-git-op.sh under PreToolUse, token-rollup-merge.sh under PostToolUse) and replace plan.md's manual execute-tally instruction with the automatic git-op recording plus roll-up reporting, so no phase is double-counted. Also drop a stray em dash from token-rollup.sh's header comment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Contributor
|
code-review-audit skipped: no audit-relevant files changed in the un-audited delta (since the last clean audit); GAIA-Audit commit status stamped on HEAD so the merge gate is satisfied with no local audit run |
Format token-rollup.sh's rendered action totals, grand total, and billing buckets with thousands separators, right-aligned into shared columns (totals to the grand-total width, buckets to the widest bucket) for readability. Ledger values and all internal arithmetic stay raw; grouping happens only at the print layer. Update the reader's bats oracles to match. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Contributor
|
code-review-audit skipped: no audit-relevant files changed in the un-audited delta (since the last clean audit); GAIA-Audit commit status stamped on HEAD so the merge gate is satisfied with no local audit run |
Contributor
|
code-review-audit skipped: no audit-relevant files changed in the un-audited delta (since the last clean audit); GAIA-Audit commit status stamped on HEAD so the merge gate is satisfied with no local audit run |
… lines (#539) A ledger line that is valid JSON but not an object (a bare scalar or array) survived the try/fromjson corrupt-line guard, then threw on .spec_id indexing and aborted the whole filter, silently dropping every good row. Coerce non-objects to the existing __BAD__ sentinel so they bump the bad count and fire the partial marker instead. Adds a bare-scalar/array fixture and a regression test. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…-accounting # Conflicts: # CHANGELOG.md
… edge (#539) The command matchers use a newline as a shell separator, so a heredoc body line that begins with the matched command does match. The comments overclaimed 'never inside a heredoc body'; correct them to note the benign, accepted edge (mid-line prose in a quoted string still does not match). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
GAIA-Audit: 1.6.1 08733ae
Contributor
|
code-review-audit skipped: GAIA-Audit trailer matches version 1.6.1 tree 08733ae |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Hardens SPEC-013's execute-time token + time tally so it:
PreToolUsehook ongit commit/git push), not a prose bullet a cold orchestrator can skip.ended_at).PostToolUsehook ongh pr merge, omitting the spec line for a spec-less plan.Phases
Phase 1 (this PR, first commit):
.gaia/scripts/token-rollup.sh— the ledger roll-up reader + dedup/aggregate/render (FC-1). 15/15 bats..claude/hooks/token-tally-git-op.sh+ shared resolver lib.claude/hooks/lib/gaia-active-plan.sh— the git-op recording hook (FC-2/FC-4). 13/13 bats.Phase 2 (follow-up commit):
.claude/hooks/token-rollup-merge.sh— the merge-time roll-up hook (FC-3)..claude/settings.json+plan.mdwiring — register both hooks, remove the manual execute tally (FC-5).Test plan
bats .gaia/scripts/tests/token-rollup.bats— greenbats .gaia/tests/hooks/token-tally-git-op.bats— greenpnpm typecheck && pnpm lint— cleancode-review-auditbefore merge.🤖 Generated with Claude Code