The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8). Its core functionality involves analyzing untrusted code and external web data using tools like Read, Grep, and WebFetch. When combined with capabilities like Bash, Write, and Edit, this creates a potential vector where malicious instructions embedded in the processed data could influence the agent to perform unintended actions.\n
Ingestion points: SKILL.md (uses Read, WebFetch, and WebSearch tools to ingest external content).\n
Boundary markers: Absent; the provided report templates do not define explicit delimiters to isolate untrusted content from system instructions.\n
Capability inventory: SKILL.md grants access to Bash, Write, Edit, and Task tools, which could be misused if the agent follows instructions found in untrusted data.\n
Sanitization: No validation or sanitization logic is present in the templates to filter malicious sequences in ingested data.

deep-analysis