octocode-skills
Pass
Audited by Gen Agent Trust Hub on Aug 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches content from external sources to discover and download new agent skills.
- Evidence: Interacts with registries such as skills.sh, aiskillstore.io, and claude-plugins.dev as documented in discovery-surfaces.md.
- [COMMAND_EXECUTION]: Employs local Node.js scripts and CLI tools to perform deterministic tasks like linting and symlinking.
- Evidence: Uses scripts/skill-review.mjs for quality checks and scripts/skill-sync.mjs for installing skills into agent platform directories.
- [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection by processing instructions from third-party skills.
- Ingestion points: Remote skill manifests and code folders fetched during discovery in fetch-remote.md and search-playbook.md.
- Boundary markers: Implements a multi-step flow requiring manual inspection and approval before installation as defined in output-format.md.
- Capability inventory: Provides the ability to write to local directories and configure lifecycle hooks through scripts/skill-sync.mjs and references/hooks-add.md.
- Sanitization: Includes a dedicated Quality Rubric (quality-rubric.md) and the skill-review.mjs linter to detect structural and safety issues in imported skills.
Audit Metadata