fix(research): require explicit invocation like the frontend skill#30
Merged
Conversation
An ordinary Codex prompt ('조사해서 정리해줘' — investigate and
summarize) activated superloopy-research: the skill declared
SUPERLOOPY RESEARCH ENABLED and began composing research lanes. Its
trigger list treated genericized phrases ('deep research', 'a request
for a cited research report, in any language') as explicit demand, and
its Codex agent metadata carried no explicit-only clause at all.
Apply the same hardening the frontend skill received for issue #28:
activation requires the explicit $/slash invocation, a leading
loopy/루피 research demand, or an active loop routing a research
deliverable here; plain research/investigate/summarize vocabulary is
named as non-activation in the skill description, the activation
section, the Codex default_prompt, and all five README rows. Contract
tests pin the explicit-only language in both metadata surfaces.
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
Synchronize the npm, Codex, and Claude manifests for the research explicit-activation fix.
Mark the banned-vocabulary list as illustrative with 'or similar', add the frontend-style negative guards so re-adding auto-activation or a trigger list fails the suite, and pin the research invocation tokens in every README locale.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
superloopy-researchactivation behind explicit invocation, mirroring the frontend skill's issue [기능 추가] 좋은 UX를 설계하는 법을 추가해주시면 감사하겠습니다. #28 hardening (merged in feat(frontend): expand cross-platform UX coverage #29)agents/openai.yamldefault_prompt, and all five README rowstest/plugin.test.jsand align the file-audit and golden-set obligationsWhy
The research skill self-activated on ordinary prompts, exactly as the frontend skill did before #28. Reproduced on a real Codex host: the prompt below — no
loopy, no "deep research", no report request — loaded the skill, printedSUPERLOOPY RESEARCH ENABLED, and began composing parallel research lanes:Two structural causes:
The prompt hook was not the leak — its triggers are already explicit-prefix only (
loopy/lpy/ultrawork).Validation
npm test— 652/652 passednpm pack --dry-run --json— version 0.11.1; 507 files