from agents.md17
Automated autoresearch loop that scans recent X (Twitter) discourse to propose, validate, and optionally apply one atomic improvement per governance file; desig
The Governance Autoresearch skill runs a focused research loop for governance documentation: it searches recent social discourse (X) for signals, cross-references bookmarks/collected data, proposes exactly one atomic change per file, validates structural and SSOT constraints, and records or commits the accepted change. The intent is safe, evidence-backed incremental improvements rather than open-ended rewrites.
Use this skill when maintainers want to keep governance files current with community practice — for example, after major ecosystem changes or periodically as part of docs maintenance. Trigger it when asked to "improve" or "audit" governance material, or when bookmarks/imported X data indicate emerging patterns.
scripts/ (has_scripts=true) to fetch candidate posts and run structural checks.Suitable for orchestrator agents that can run Python scripts and access X (Twitter) via a bearer token; examples: Codex/Gemini CLI-driven automation, Python-capable agent runtimes, or CI jobs that apply documentation edits after validation.
Governance autoresearch skill searches X (Twitter) for recent discourse on governance topics and proposes atomic improvements. The bundled script crashes immediately due to hardcoded path assumptions (parents[3]) and a dependency on a local x_runtime module that only exists in the source repo. SKILL.md workflow is well-documented but the script is non-portable.
x_runtime (local module from parmartejass/agents.md repo)Not malicious but essentially non-functional outside its original repo. The x_runtime dependency and hardcoded path structure make it a repo-specific utility rather than a portable skill. The concept (autoresearch loop for governance improvement) is interesting but execution is tightly coupled to one repo's layout.
