The Agent Skills Directory

[SAFE]: The skill is a privacy auditing tool. It uses a local Python script (scan_publication_candidates.py) to identify and redact sensitive information such as API keys, private IPs, and email addresses. The script is self-contained and does not perform network operations or install external dependencies.- [PROMPT_INJECTION]: The skill processes untrusted user-supplied content (private files), creating a surface for indirect prompt injection. The ingestion occurs via the scan_publication_candidates.py script. The skill mitigates this risk by redacting sensitive evidence and HTML-escaping output in its generated reports. Furthermore, the instructions in SKILL.md and references/classification-policy.md establish a conservative 'private-only' default for the agent's decision-making process.

memory-publication-auditor