The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill creates a potential surface for indirect prompt injection as it ingests untrusted data from both user arguments and the local codebase to perform its analysis. This is a functional requirement of its primary purpose.
Ingestion points: Processes user input via $ARGUMENTS and codebase content via Read, Grep, and Glob tools.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are included in the subagent prompt in SKILL.md.
Capability inventory: The skill has filesystem write access via the Write tool and shell execution via restricted Bash commands (git and node).
Sanitization: No validation or filtering is applied to the ingested content before it is processed by the agent.
[DATA_EXPOSURE_AND_EXFILTRATION]: No patterns of accessing sensitive credentials (like .ssh or .env files) or exfiltrating data to external domains were identified.
[UNVERIFIABLE_DEPENDENCIES]: The skill does not perform external package installations or download scripts from remote servers.

deep-analyze