ci: block backward-framing tell-words in the guard#109
Merged
Conversation
The guard caught string leaks (codename, eval/corpus tokens) but not the conceptual leak that slipped into the boundaries doc — prose implying an external consumer finds agents 'confusing'. Adds a second category of blocked phrases (commonly confused / easy to confuse / trap word / seem(s) to overlap) with its own 'Backward-framing tell-word' error, steering authors to forward ownership framing. Same internal-ref-allow escape hatch. Verified: clean=pass, both categories fail with distinct labels, allow-marker ignored. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Hardens
check-no-internal-refs.shto catch the conceptual leak class the string-grep missed.Why
The guard blocked
tarmac/eval-runs/routing_truth, but a boundaries doc still shipped backward-framing prose ("commonly confused", "trap word", "seem to overlap") — which implies an external consumer finds the agents confusing. That is moat-adjacent even with zero internal strings, and it took a human re-read to catch. This turns that into a permanent rule.What
Backward-framing tell-wordcategory (distinct error from internal references):commonly[ -]confused,easy to confuse,trap word,seem(s) to overlap.internal-ref-allowper-line escape hatch.Tested
Docs/CI only.