Nm Memory Palace Knowledge Intake

Process external resources into stored knowledge with quality evaluation, curation routing, and application decisions

This skill is coherent for its stated goal of ingesting and curating external resources, but it assumes runtime tools and remote conversion endpoints that are not declared in the registry metadata. Before installing/use: (1) Confirm where the required 'leyline' document-conversion and content-sanitization hooks point — if they call remote services, do not send sensitive local files until you trust the endpoint. (2) Be aware the SKILL.md will construct file:// URIs for local file ingestion — only supply local paths you are comfortable having processed. (3) The discussion-promotion module uses the gh CLI to create/update GitHub Discussions; ensure the agent's gh auth and repository permissions are appropriate and you consent to posting summaries. (4) Ask the skill author to explicitly list required binaries (gh, markitdown or conversion tools, WebFetch) and any credentials or endpoints the hooks call. (5) If you cannot verify the document-conversion and WebFetch implementations are local/trusted, test in a sandbox environment first or decline to enable automatic ingestion of private files. If the author supplies clarifications showing all conversion and storage endpoints are local or under your control, this skill is reasonably coherent; otherwise treat it cautiously.

Type: OpenClaw Skill Name: nm-memory-palace-knowledge-intake Version: 1.0.0 The skill bundle provides a comprehensive framework for an AI agent to ingest, evaluate, and manage external knowledge resources. It includes instructions for fetching web content, performing quality validation via a 'slop-detector', and promoting high-quality entries to GitHub Discussions using the 'gh' CLI (modules/discussion-promotion.md). While the skill involves high-privilege actions such as network access and GitHub API interactions, these behaviors are transparently documented, strictly aligned with the stated purpose of knowledge management, and emphasize human-in-the-loop approval for any destructive or public-facing actions (modules/pruning-workflows.md). No evidence of malicious intent, data exfiltration, or unauthorized persistence was found.