no-comments
Pass
Audited by Gen Agent Trust Hub on Nov 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates entirely within the project scope provided by the caller to identify and suggest the removal of unnecessary comments. It includes human-in-the-loop checkpoints for critical changes and relies on standard agentic patterns (
/swarm,/architect,/why). - [EXTERNAL_DOWNLOADS]: The skill references external sources in
UPSTREAM.md(pointing to thecursor/pluginsrepository), but these are for documentation and provenance tracking rather than runtime code execution. The downloads are from a well-known service (GitHub) and a recognized community repository, which is considered safe. - [COMMAND_EXECUTION]: The skill uses internal agent commands such as
/swarm,/architect, and/why. These are standard tool invocations within the agent's environment and do not involve arbitrary shell execution or unsafe user-input interpolation. - [INDIRECT_PROMPT_INJECTION]: The skill processes code and diffs, which are technically untrusted data. However, it implements strict boundary markers and logic (Step 2 in
SKILL.md) to verify findings, trace symbols, and reject 'application-code edits' or 'scope escapes'. The instructions explicitly tell the agent not to write application code, limiting the impact of potential injections to comment deletions only.
Audit Metadata