The Agent Skills Directory

[PROMPT_INJECTION]: The skill provides structured instructions for a multi-stage development pipeline (advisor, probe, ship). No override markers, bypass attempts, or adversarial role-playing instructions were found.
[DATA_EXFILTRATION]: There are no network operations, hardcoded credentials, or access requests for sensitive environment variables or configuration files. The skill operates locally on the current worktree.
[REMOTE_CODE_EXECUTION]: The skill does not include package installation commands or remote script fetching. Its primary function is generating boilerplate code stubs (typed signatures with empty bodies) for human review.
[OBFUSCATION]: The skill instructions are written in plain markdown. No hidden encoding, zero-width characters, or steganographic techniques were detected.
[INDIRECT_PROMPT_INJECTION]: While the skill ingests data from external sources like tickets or previous conversation logs, the risk is mitigated because the output is restricted to non-executable code stubs meant for interactive user review.

probe