adhd-body-doubling

Pass

Audited by Gen Agent Trust Hub on Mar 4, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes a local helper script scripts/start-session.sh used to initialize focus sessions. This script collects user input (tasks, energy levels) and generates a JSON history file. The script utilizes standard shell practices and includes a json_escape function to ensure that user-provided strings are correctly formatted before being written to the data files.\n- [PROMPT_INJECTION]: The skill implements a session history feature (Category 8 surface) where past task descriptions are stored and can be reviewed by the agent to identify productivity patterns. \n
  • Ingestion points: Session history is read from ~/.openclaw/skills/adhd-body-doubling/history/ when the user requests a history overview.\n
  • Boundary markers: The agent is guided by strict behavioral protocols defined in SKILL.md and references/protocols.md which specify how history data should be processed.\n
  • Capability inventory: The agent can execute the provided local initialization script and read/write to the skill's specific history directory.\n
  • Sanitization: User input is escaped via the json_escape helper in the shell script before persistence, reducing the risk of data being misinterpreted as commands or instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 4, 2026, 09:28 PM