roadmap
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external GitHub and GitLab sources, including issue titles, comments, and pull request descriptions (
references/competitive-scan.md). This creates a surface for indirect prompt injection where an attacker could place instructions in a public repository that the agent might follow when performing a competitive scan.\n - Ingestion points: External repository READMEs, Issues, PRs, and Discussions.\n
- Boundary markers: The skill writes to a structured
ROADMAP.mdfile but does not explicitly use delimiters or ignore-instructions warnings for the external content.\n - Capability inventory: Executes shell commands (
git,gh,glab) and writes to local files (ROADMAP.md,.gitignore).\n - Sanitization: Implements a signal-based fit filter and requires explicit user approval before adding findings to the roadmap.\n- [COMMAND_EXECUTION]: The skill invokes several shell commands to manage project state and fetch remote data. It uses
git log,git tag,gh pr list, andglab mr listacross various workflows (SKILL.md,references/competitive-scan.md).
Audit Metadata