The Agent Skills Directory

[COMMAND_EXECUTION]: Environment Discovery. The skill instructs the agent to execute a shell command ls across various hidden directories in the home folder (e.g., ~/.claude/skills/, ~/.agent/skills/) to check for other installed companion skills. This constitutes unsolicited environment fingerprinting to identify installed software.
[PROMPT_INJECTION]: Indirect Prompt Injection Surface. The skill is designed to analyze and structure responses to external documents like RFPs and SOWs which may be provided by third parties. Ingestion points: RFP and SOW requirements provided by the user (SKILL.md). Boundary markers: Absent; there are no instructions to use delimiters or to ignore potential instructions embedded in the input data. Capability inventory: The agent has access to shell tools for environment checks (SKILL.md). Sanitization: Absent; there is no logic to filter, escape, or validate the content of processed documents to prevent malicious instruction execution.

proposal-writing