serp-analysis
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external search results and third-party web pages, creating a potential surface for indirect prompt injection attacks. \n
- Ingestion points: The skill retrieves data through the
WebFetchtool and custom connector scripts (firecrawl.pyandtavily.py) as specified in theData Sourcessection ofSKILL.md. \n - Boundary markers: The instructions include a clear security boundary warning that
WebFetchcontent is untrusted and should be treated as evidence only. \n - Capability inventory: The skill is capable of network access via
WebFetchand local shell execution usingpython3. \n - Sanitization: The instructions direct the agent to flag any prompt-like directives or owner overrides found in fetched content and avoid following them. \n- [COMMAND_EXECUTION]: The skill instructions direct the agent to execute local Python scripts to interact with search engines. \n
- Evidence: The
Data Sourcessection inSKILL.mdprovides shell commands for the agent to run, such aspython3 "${CLAUDE_PLUGIN_ROOT}/scripts/connectors/firecrawl.py"andtavily.py.
Audit Metadata