The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external data sources such as session logs, artifacts, and commit history as described in SKILL.md.
Ingestion points: The skill reads logs, artifacts, commits, and verification results from the current or specified previous sessions.
Boundary markers: The instructions do not define clear delimiters or "ignore embedded instructions" warnings when the agent processes these external data sources.
Capability inventory: The skill identifies and proposes fixes for repository configuration, automated checks (build scripts, CI, hooks), steering files, and tool use based on its analysis.
Sanitization: There is no mention of sanitizing, escaping, or filtering the content found in logs or artifacts before the agent interprets them.
Security Impact: While the skill emphasizes evidence-based reasoning, a malicious actor who can influence log output or commit messages (e.g., via a Pull Request or a compromised test run) could attempt to trigger an indirect prompt injection to mislead the agent into recommending insecure configuration changes.

retro