agentleFS
emergent-wisdomemergent-wisdom/understanding-graph/skills/quality-check/SKILL.md

quality-check

Skill1 starsChanged 7 months agoMITView on GitHub

Inspect graph structure when the user asks for a check or when a concrete structural problem may be affecting the work. Treat metrics as diagnostics, never as quotas or semantic-quality scores.

10 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill deliberately, not on a timer.

    line 22
  2. Very specific:

    When a finding could materially affect the task, use graph_suggest_next({ task, workflow }) for concrete weighted routes or …

    line 55
  3. Distinctive:

    If the diagnosis identifies a genuine issue, make the smallest truthful change:

    line 63
  4. Distinctive:

    leave healthy heterogeneity alone.

    line 70

“Common” is measured: kept by at least 6 in 10 of the 8 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Quality Check

Use this skill deliberately, not on a timer. Do not interrupt useful work for scheduled maintenance, demand trigger diversity, or ask the user to repair the graph merely because a number is low.

Diagnose before changing

Choose the smallest view that answers the live question:

  • graph_analyze({ include: ["gaps", "bridges", "questions"] }) for islands, bridges, and genuinely unanswered questions;
  • graph_score() for a compact structural snapshot;
  • graph_centrality() when influence or bottlenecks matter;
  • graph_semantic_gaps() when embedding coverage exists and conceptual distance is relevant;
  • graph_skeleton() only when the overall topology is itself in question;
  • graph_thermostat() as an advisory entropy-style pulse, not a command.

Metrics describe the stored shape. They cannot establish truth, creativity, importance, completeness, or whether a node was worth creating. A sparse graph can be exactly right for a small task; a dense graph can still be confused.

Judge the finding in context

  • A disconnected region is a problem only if the work gives it a real relation to another region. Do not invent a bridge.
  • An unanswered question may remain intentionally open.
  • A contradiction is live conflict, not proof that either side is false.
  • A low supersession count is not evidence that someone failed to reconsider.
  • Trigger counts are descriptive. Never add questions, tensions, predictions, or any other type to improve a distribution.
  • Artifact-heavy structure is an advisory signal only. Add cognitive testimony only when genuine understanding, uncertainty, evidence, or choice exists.

When a finding could materially affect the task, use graph_suggest_next({ task, workflow }) for concrete weighted routes or graph_discover_grounded() for a bounded distant comparison. Choose, modify, combine, or reject the result. “No defensible connection” and “no repair needed” are valid outcomes.

Repair only what is real

If the diagnosis identifies a genuine issue, make the smallest truthful change:

  • connect regions with a specific typed relation;
  • preserve an unresolved tension rather than forcing consensus;
  • answer a question only when the graph now contains an answer;
  • revise or use dedicated graph_supersede when a position actually changed;
  • archive noise through the explicit lifecycle operation;
  • leave healthy heterogeneity alone.

Explain the material finding to the user when it changes the work or needs their judgment. Otherwise keep the check quiet and continue the task.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.