chart-assistant
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill's primary instruction file (SKILL.md) contains explicit defensive instructions directing the AI agent to ignore instructions or commands embedded within user-provided data, effectively guarding against malicious attempts to override agent behavior.
- [DATA_EXFILTRATION]: The skill establishes strict data handling rules in SKILL.md and existing-card-replacement.md, explicitly forbidding the AI from outputting, saving, or transcribing sensitive information such as authentication tokens, app secrets, or document signatures.
- [EXTERNAL_DOWNLOADS]: Documentation and installation instructions reference official and well-known platforms, including Feishu Open APIs and Vercel's GitHub-hosted skills CLI, ensuring that all external interactions occur with established services.
- [INDIRECT_PROMPT_INJECTION]: The skill handles untrusted data from sources like spreadsheets and URLs as described in workflow.md and sources-and-sync.md. It mitigates injection risks by implementing boundary markers (instructing the agent to treat input as data, not instructions) and by employing structural sanitization via a local validation script (scripts/common-option-validation.mjs). The skill's capabilities are limited to authorized document updates via the Feishu API as documented in feishu-doc-card.md.
Audit Metadata