The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data access were detected. The skill correctly implements least-privilege principles by restricting file write operations to a dedicated 'inbox/' directory structure.
[PROMPT_INJECTION]: The skill processes untrusted external data such as meeting transcripts, chat exports, and user-provided links, which represents a surface for indirect prompt injection.
Ingestion points: External files and links processed via the 'Batch Ingest' and 'External Source' capture modes in SKILL.md and references/batch-ingest.md.
Boundary markers: No explicit delimiters or instructions are provided to the agent to treat external content as untrusted data.
Capability inventory: File system access is limited to writing new files in inbox/ and raw/, and appending to activity logs in log/.
Sanitization: No explicit sanitization or filtering of external input is described.

pk-capture