mission-control

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow where task descriptions are retrieved from an external API (https://claw-controlbackend-production.up.railway.app/api/tasks) and passed to the sessions_spawn tool to create new agent environments. Ingestion points: Data enters the context via the /api/tasks API endpoint. Boundary markers: The skill does not define specific delimiters or instructions to sub-agents to ignore instructions embedded in the task data. Capability inventory: The skill uses sessions_spawn for agent creation, curl for network access, and expects agents to perform git operations. Sanitization: No sanitization or content validation is performed on the task data before it is relayed to sub-agents.
  • [DATA_EXFILTRATION]: The skill transmits project-related metadata, including detailed task descriptions, agent activity logs, and commit hashes, to an external backend at claw-controlbackend-production.up.railway.app. This enables the task tracking functionality but results in project data leaving the local environment.
  • [COMMAND_EXECUTION]: The agent is instructed to use the curl command-line tool to perform GET, POST, and PUT requests to interact with the external Mission Control API.
  • [EXTERNAL_DOWNLOADS]: The skill fetches task lists and message feeds from the external API endpoint hosted on the Railway platform.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 05:12 AM
Security Audit — agent-trust-hub — mission-control