agent-lynx
Pass
Audited by Gen Agent Trust Hub on Nov 18, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads and executes the
agent-lynxCLI package from the NPM registry usingnpx --yes agent-lynx@latestor via globalnpm install. It also downloads a native TASM codec backend binary during the first execution of conversion tasks, which is verified using SHA-512.\n- [DYNAMIC_EXECUTION]: The skill employs a dynamic instruction-loading mechanism. It discovers and fetches task-specific sub-skills at runtime via theagent-lynx skills getcommand or by readingSKILL.mdfiles and associated resources directly from thelynx-community/skillsrepository on GitHub.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external developer artifacts, creating an ingestion surface for untrusted data.\n - Ingestion points: Performance traces, recorder replay JSON, recorder fixtures, and TASM templates.\n
- Boundary markers: The instructions do not specify any delimiters or safety markers to isolate the content of these artifacts from the agent's instructions.\n
- Capability inventory: The skill possesses the capability to execute shell commands (via the CLI tool) and read/write files on the local system.\n
- Sanitization: No sanitization or validation of the data within the processed artifacts is described.
Audit Metadata