The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill relies on external files to define its execution protocol, creating an ingestion surface where untrusted content could influence agent behavior.
Ingestion points: The skill reads from ../alignfirst/SKILL.md and references/catchup-protocol.md.
Boundary markers: Absent. There are no markers or instructions to isolate or ignore potentially malicious content within the referenced files.
Capability inventory: The skill directs the agent to "Execute the catchup protocol," which may involve tool usage or complex logic defined in the external files.
Sanitization: Absent. The skill does not perform any validation or filtering of the content loaded from the external references.
[PROMPT_INJECTION]: The skill contains instructions that explicitly override the agent's standard operating procedures.
Evidence: The directive "Do not use your own plan mode" modifies the agent's reasoning process and autonomy.

alcatchup