The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from project thread history to identify user patterns and preferences.
Ingestion points: Step 1 (Mine their history) describes using t3_thread_read to survey recent agent conversations across history slices.
Boundary markers: No specific delimiters or warnings to ignore embedded instructions in the thread history are specified, potentially allowing instructions in the ingested data to influence the mining process.
Capability inventory: The skill possesses file-writing capabilities (via skill-creator) and project management capabilities (via gh pr create and link_pull_request).
Sanitization: No sanitization or validation of the mined thread content is performed before it is used to author the final -mode skill draft.

automate-me