The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from multiple sources including git logs, commit diffs, CI job logs, and dependency lockfiles. While it uses these to generate a report, malicious content in these sources could potentially influence the agent's behavior during the analysis phase.
Ingestion points: Git log, diff hunks, CI logs (gh run view --log-failed), dependency manifests (package.json, pyproject.toml, etc.), and benchmark artifacts.
Boundary markers: Instructions require placing logs in code blocks and separating "observed" vs "suspected" findings.
Capability inventory: Access to shell via Bash tool, reading files, and performing web fetches.
Sanitization: No explicit sanitization or filtering of log content or diffs is mentioned before processing.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute various development and quality assurance utilities.
Evidence: Executions include git log, gh run list, npm outdated, pip list --outdated, cargo outdated, and potentially infrastructure scanners like tfsec or hadolint if triggered. These are aligned with the skill's purpose.

daily-qa