clawhub-batch-analysis
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes third-party plugin code exported from ClawHub, which constitutes a potential attack surface for instructions embedded in the analyzed data.
- Ingestion points: Plugin ZIP exports retrieved from the
/api/v1/plugins/exportendpoint as described in SKILL.md. - Boundary markers: The instructions lack explicit structured delimiters for the ingested plugin data, though they include prose warnings for the agent to treat the source as untrusted.
- Capability inventory: The skill leverages the
plugin-inspector batchcommand to perform analysis on the exported files. - Sanitization: The skill contains explicit safety notes directing the agent not to use execution-enabling flags like
--runtime,--mock-sdk, or--allow-executeunless in an isolated workspace. - [COMMAND_EXECUTION]: The skill involves the execution of a CLI utility to process the exported plugin corpus.
- Evidence: Instructions to run the
plugin-inspector batchcommand with specific flags for output management and concurrency in SKILL.md.
Audit Metadata