make-and-share-skill
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external social media posts to create executable instructions for agents.\n
- Ingestion point: Step 1 instructions describe reading posts from X via a plugin.\n
- Capability inventory: The skill utilizes github-mcp-repo-writes-enabled to commit code and tools.owl to perform automated browser actions in a signed-in session.\n
- Boundary markers: No explicit delimiters or instructions are provided to separate data from commands or to ignore instructions embedded within the source content.\n
- Sanitization: There is no validation or sanitization to prevent adversarial instructions in the source post from being executed or included in the generated output.\n- [DYNAMIC_EXECUTION]: The skill dynamically constructs a new SKILL.md file at runtime using contents derived from an untrusted external source and pushes it to a public GitHub repository, which other users are encouraged to install.
Audit Metadata