The Agent Skills Directory

[SAFE]: The skill processes 'thought' inputs provided by the agent. Ingestion points occur via the '--thought' argument in 'scripts/process-thought.js' and 'scripts/format-thought.js'. Boundary markers are provided by the methodology instructions in 'SKILL.md' which structure the agent's reasoning. The capability inventory is limited to local file system writes for history tracking and console output for display, with no network or shell execution access. Sanitization is performed by treating all inputs as plain text and using standard JSON stringification for storage.- [SAFE]: No external network requests or remote code downloads were identified. The logic is entirely contained within the provided local scripts.- [SAFE]: No hardcoded secrets, sensitive file paths, or administrative privilege requirements were detected in the source code or configuration files.

sequential-thinking