pipeline-review
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes data from external CRM systems and databases.\n
- Ingestion points: Deal and meeting data are extracted from Salesforce, HubSpot, Supabase, Notion, and CSV files as described in Step 1 of SKILL.md.\n
- Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard potential commands hidden within the ingested sales data.\n
- Capability inventory: The skill has the ability to write reports to the local file system and send summaries to external platforms like Slack and via email (Step 4, SKILL.md).\n
- Sanitization: The instructions do not include steps to sanitize or validate the content of fields such as 'Deal name' or 'Loss reason' before processing.\n- [COMMAND_EXECUTION]: The skill meta-information specifies an installation command using 'npx goose-skills install pipeline-review' (skill.meta.json). This executes a CLI tool to manage skill installation, which is a standard procedure for this vendor's ecosystem.\n- [DATA_EXFILTRATION]: The skill handles sensitive sales and pipeline data, including revenue amounts and account names. It writes this information to local files and can export data to external services such as Google Sheets, Slack, and email (SKILL.md). These features are consistent with the tool's purpose for reporting and collaboration.
Audit Metadata