agentleFS
Pearl1531Pearl1531/Archetype-talk/.claude/skills/contradictions/SKILL.md

contradictions

Skill2 starsChanged 50 days agoMITView on GitHub

Find and report places where the graph disagrees with itself — Signals vs Signals, Signals vs Evidence, personas vs their sources. Use when the user runs /contradictions, asks where the data conflicts, or when another skill flags a divergence.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Don't manufacture tension — near-misses and different-but-compatible observations are not contradictions.

    line 37

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “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

contradictions — Where the data disagrees with itself

"Contradictions are data" is a standing rule of this repo — but until now there was no view that collects them. This skill reads the graph semantically and reports every place it argues with itself. Report-only: it never resolves anything.

Trigger

/contradictions                    ← whole graph
/contradictions Personas/Emma.md   ← one persona's slice

What counts as a contradiction

  1. Signal vs Signal — two participants (or the same one across sessions) reporting opposite behaviour/attitude on the same thing.
  2. Signal vs Evidence — an interview observation that runs against external data (e.g. a participant loves a feature the community data says is broadly hated — or vice versa).
  3. Persona vs sources — a Pain, quote, or Product usage line in a Persona file that its own linked Signals no longer support (drift after edits).
  4. Correlation vs new data — a ## Correlations pattern that a newer Signal/Evidence undercuts.
  5. Tone drift — canonical Tone of Voice traits vs contradicting candidate observations (extract-findings flags these at import; this skill re-surfaces unresolved ones).
  6. Stance split — two Signals recording opposite sentiment.stance values on the same feature:. Cheap to find (grep -A2 '^sentiment:' Signals/*.md) and easy to under-read, so treat the two ends as the interesting cases: a dealbreaker next to a relies_on is a segment boundary or a change over time, not noise to be split down the middle. Report the two stances with their because: lines side by side. Never average, and never convert the ladder to numbers to find a midpoint — a persona whose stances got reconciled is a persona nobody can act on. mixed inside a single signal is not a contradiction; it is one person being genuinely ambivalent, which is a finding in its own right.

Report format

Per finding:

  • The tension in one sentence — plain language.
  • Both sides, cited — file names + the exact quotes/values that clash, each with its Level (analyst context, so codes like L3 are fine here).
  • Possible readings — e.g. different segments, different contexts, time drift, or a genuinely open question. Never pick a winner.
  • Suggested next step — usually a Research backlog.md row ("which root cause is it?"); offer to append it.

Rules

  • Never resolve a contradiction by editing files. Two conflicting Signals stay two files with different root causes. If the user wants to act, route to /backlog (research it) — the answer comes from real users, not from us.
  • Don't manufacture tension — near-misses and different-but-compatible observations are not contradictions; when in doubt, leave it out.
  • A clean report ("no contradictions found") is a valid, useful result — say it plainly.

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.