fnl-open-source-review

Pass

Audited by Gen Agent Trust Hub on Nov 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute various command-line tools for auditing purposes, such as git, gh api, gitleaks, and trufflehog.
  • Evidence: The instructions specify using git ls-files, git log -p, and gh api to inspect repository state and settings.
  • Context: These tools are used in a read-only manner for auditing. The skill explicitly advises against the agent installing software itself, recommending instead that the user perform installations.
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves reading and analyzing arbitrary content from external repositories, which presents a surface for indirect prompt injection if those files contain malicious instructions.
  • Ingestion points: The agent reads the entire working tree and git history of a specified repository (SKILL.md).
  • Boundary markers: None explicitly defined to separate audited content from agent instructions, though the task context is strictly defined as a "read-only review".
  • Capability inventory: The skill utilizes gh api for metadata retrieval and git for history/file inspection. It does not perform file writes or network exfiltration of the data read.
  • Sanitization: No specific sanitization of the ingested repository content is mentioned, though it instructs the agent to report findings rather than execute code found within the repository.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 4, 2026, 03:19 PM