skills/sanjay3290/ai-skills/manus/Gen Agent Trust Hub

manus

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes output from an external service, creating a surface where content from the Manus API could influence agent behavior. (1) Ingestion points: Data enters the agent context through API responses from the GET /v1/tasks endpoint in SKILL.md. (2) Boundary markers: The instructions lack explicit delimiters or warnings to treat API output as untrusted data. (3) Capability inventory: The skill relies on curl for network operations and jq for parsing, and is intended for use with an agent capable of multi-step workflows. (4) Sanitization: API responses are processed via jq to extract text and file URLs without additional validation or filtering.
  • [EXTERNAL_DOWNLOADS]: The skill fetches task metadata, reports, and file attachments from the Manus AI domains api.manus.ai and manus.im.
  • [COMMAND_EXECUTION]: The skill uses curl and jq commands to interact with the Manus AI service as documented in SKILL.md and references/api.md.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 01:57 PM
Security Audit — agent-trust-hub — manus