The Agent Skills Directory

[SAFE]: The skill implements standard development workflows for validating task completion, including test execution and build verification.
[COMMAND_EXECUTION]: Executes standard toolchain commands such as npm test, pytest, cargo test, and go test based on the project environment detected via manifest files.
[DATA_EXPOSURE]: Accesses project manifest files (e.g., package.json, Cargo.toml) and repository state via git stash to evaluate build readiness and test coverage regressions.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from task descriptions and review comments.
Ingestion points: task.description and reviewResults in SKILL.md.
Boundary markers: None explicitly implemented within the extraction logic.
Capability inventory: Shell command execution (npm/pytest/cargo/go) and file system writes via fs.writeFileSync.
Sanitization: Employs regular expressions to extract specific patterns (bullet points and numbered lists) from descriptions, which limits the influence of injected instructions.

validate-delivery