The Agent Skills Directory

[SAFE]: The skill's primary function is to summarize historical data from authorized sources. No evidence of malicious intent, credential theft, or unauthorized network activity was found.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface due to its core function of ingesting data from external sources like tickets, team discussions, and historical logs.
Ingestion points: SKILL.md (Step 3: Mining authorized working history; Step 4: Sweeping the shared record including tickets and documents).
Boundary markers: Absent. There are no instructions to the agent to use specific delimiters or to disregard instructions found within the ingested historical data.
Capability inventory: The skill utilizes tools for repository state verification, history synthesis, and multi-agent coordination via the pstack contract.
Sanitization: Absent. While the skill mentions removing private context, it does not specify sanitization or escaping of the ingested content to prevent instructions in those sources from influencing the agent's behavior.

recall