The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted source code, environment files, and dependency manifests as part of its auditing routine, which creates a vulnerability surface for indirect prompt injection.
Ingestion points: The skill reads local source code, environment variables, Git history, and dependency manifest files during the static and configuration check phases (SKILL.md, references/attack-surface.md).
Boundary markers: The instructions lack explicit requirements for the agent to use delimiters or "ignore" markers when processing the audited content to prevent instruction leakage.
Capability inventory: The agent has the capability to read/write files and execute system-level commands to perform audits and apply fixes (SKILL.md).
Sanitization: No specific sanitization, escaping, or filtering of the audited content is mandated before the agent processes the data.
[COMMAND_EXECUTION]: The skill utilizes system-level commands to perform dependency audits.
The agent is instructed to run external auditing tools such as npm audit, pip-audit, and cargo audit to identify vulnerabilities in the project's dependency tree (SKILL.md).

harden