implement-with-sol

Pass

Audited by Gen Agent Trust Hub on Sep 1, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from GitHub issues via the gh issue view command and uses it as a specification for implementation. An attacker could craft a malicious issue body to influence the behavior of the downstream codex exec command.
  • Ingestion points: The skill reads issue metadata (title and body) from GitHub and writes it to a temporary file in SKILL.md step 2.
  • Boundary markers: It uses file path referencing (e.g., /tmp/<slug>.md) to avoid direct prompt inlining, which provides a structural boundary but does not sanitize the content.
  • Capability inventory: The skill possesses significant capabilities including file system writes (codex exec --sandbox workspace-write), version control operations (git wt, git-hunk skills), and PR creation (make-pr).
  • Sanitization: No explicit sanitization, filtering, or validation of the GitHub issue content is performed before it is passed to the implementation tool.
  • [COMMAND_EXECUTION]: The skill uses various CLI tools such as git, gh, and codex to manage worktrees, fetch issue data, and execute implementation tasks. These commands are integral to the skill's core functionality for developer workflow orchestration.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 1, 2026, 01:59 AM
Security Audit — agent-trust-hub — implement-with-sol