volcengine-feedback
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the conversation history and CLI error logs to generate feedback drafts.\n
- Ingestion points: Conversation context and CLI error logs are used as source material in
SKILL.md.\n - Boundary markers: The skill is instructed to provide a compact preview for user verification before any action.\n
- Capability inventory: The skill may utilize a GitHub integration or browser to submit data as described in
SKILL.md.\n - Sanitization: Instructions explicitly mandate the redaction of sensitive data, including AK/SK, tokens, passwords, and PII, replacing them with placeholders like
<account-id>.\n- [SAFE]: All external references point to official GitHub repositories under thevolcengineorganization, which matches the skill author and intended purpose.
Audit Metadata