The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute build, test, and linting scripts found within the repository being audited as part of the verification process. This behavior is documented in SKILL.md (Step 10) and references/verification.md, allowing for arbitrary command execution based on the contents of the target project.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the repository being reviewed, including documentation, README files, and landing pages, which could contain malicious instructions designed to influence the agent's behavior during the review process.
Ingestion points: SKILL.md (Step 2: Read relevant project documentation), references/documentation-drift.md (README, docs, landing pages, API documentation).
Boundary markers: Absent; the instructions do not specify the use of delimiters or specific commands to ignore instructions embedded within the files being audited.
Capability inventory: SKILL.md (Step 10: Run available tests, lint, typecheck, and builds), references/verification.md (Run the project's existing checks).
Sanitization: Absent; there are no instructions for sanitizing or validating the content of the files before processing or executing associated scripts.
[DYNAMIC_EXECUTION]: The instructions require the agent to run existing project checks and builds (references/verification.md), which involves the dynamic execution of scripts or binaries defined by the target repository rather than static commands.

app-review