Repository navigation
Develop - #3
Merged
Merged
Develop#3
Conversation
Score your real local Claude Code sessions with zero new API calls: session-log adapter, scan subcommand, --all-projects aggregation. Drop synthetic/unknown-model sessions; tighten no-print rule. Adapter tests + README leads with scan.
Derive deterministic checks from a rules file via a pattern library (no model). Reports which rules had no auto-match. Tests + README.
- Three-valued checks (pass/fail/unknown): rejected or unconfirmed tool calls are excluded from pass-rate denominators instead of counting as executed. - Claude Code importer joins tool results (status, exit codes), dedupes streaming usage per message, labels mixed-model sessions, and no longer lets synthetic notices replace the final answer. - scan skips unreadable/invalid session files with per-file warnings (--strict stops instead); audits stay strict. - keeptrue audit prepare/report: frozen, hashed snapshots; blank reference labels; precision/recall with abstentions counted as misses. - experiments/02: retrospective pilot on one real session that exposed and fixed two importer false passes. - Report shows n/? per cell; demo hero aligned with the report's wording.
Diff evidence showed the change's first line instead of the line that matched (e.g. "+def handler(...)" for a print() violation). Evidence now quotes the matching line. Regenerate docs/demo.svg from the current report and align README captions and the no-print pattern with the tool.
…mands - keeptrue.adapters.codex reads ~/.codex/sessions rollouts. Newer rollouts use CommandExecution/FileChange records (argv, exit code, status, per-file diffs); older ones use exec_command output and apply_patch envelopes. Code-mode JavaScript never counts as a confirmed command on its own. Tokens come from per-response usage records, or the cumulative token_count total in older rollouts. - scan --agent claude|codex|all. Codex sessions are matched to a repo by their recorded working directory. The log format is detected per file, so --logs and audit prepare accept either agent. - Command checks ignore heredoc bodies fed to non-shell programs. A first cross-agent scan got three Codex cells wrong because analysis scripts quoted rule patterns inside heredocs (experiments/03). - Shared JSONL helpers move to adapters/_jsonl.py.
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
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.
No description provided.