skill-scout
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks as it is designed to ingest and read content from untrusted external sources like GitHub repositories and web search results.
- Ingestion points: External
SKILL.mdfiles retrieved via theghCLI and web search results (Step 3 and Step 4). - Boundary markers: None; the agent is instructed to read the content of the external files directly without specific delimiter-based isolation.
- Capability inventory: The skill utilizes
findandgrepfor local filesystem inspection andghfor remote repository searching (Step 2 and Step 3). - Sanitization: The skill explicitly includes a security vetting procedure (Step 4
- Vet External Matches) that instructs the agent to manually inspect external instructions for malicious patterns, such as unexpected shell commands, network calls, or credential handling, which serves as a procedural mitigation against injection attacks.
Audit Metadata