The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized operations were detected.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes user-supplied notes. 1. Ingestion points: Material is read from member workspaces and conversation summaries (SKILL.md). 2. Boundary markers: No explicit delimiter markers are used for untrusted data. 3. Capability inventory: The skill has the capability to write files to the project knowledge directory (SKILL.md). 4. Sanitization: It instructs the agent to filter out command chatter and private notes. The risk is significantly reduced by a mandatory dry-run review before any files are written.
[DATA_EXFILTRATION]: The skill includes guardrails preventing the storage of secrets, credentials, or private data, and lacks network capabilities.

knowledge-capture