agentleFS
alekspetrovalekspetrov/navigator/skills/nav-triz/SKILL.md

nav-triz

Skill356 starsChanged 13 months agoMITView on GitHub

Divergent solving for a declared contradiction ("improving X worsens Y"). Produces three competing candidates from different TRIZ separation modes, each with cost and downside, then recommends one. Auto-invoke when user says "find a better solution", "alternatives for", "resolve the contradiction", "triz this", or when a nav-brief declares a Contradiction on a substantial task.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    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. Very specific:

    Does NOT fire for routine work, a brief with Contradict none, "just do it" / "quick fix" passthroughs, or when the user already named the …

    line 22

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Navigator TRIZ Skill

Navigator finds a solution fast. This skill exists for the minority of tasks where the first fitting idea is a compromise: improving one thing worsens another. It forces three genuinely different candidates onto the table before PLAN picks one, then records the resolution so the next session can find it.

When This Fires (and when it must not)

Fires when either:

  • The user asks for it: "find a better solution for X", "alternatives for X", "resolve the contradiction", "triz this".
  • A nav-brief Contradict row is not none AND the task is substantial (Task Mode complexity ≥ 0.5, or touching more than one subsystem).

Does NOT fire for routine work, a brief with Contradict none, "just do it" / "quick fix" passthroughs, or when the user already named the approach. Ceremony is the failure mode of this skill; when in doubt, skip it and say so in one line. (Resolved before: mem-060, mem-074 — new blocking behaviour ships opt-in, by condition.)

Protocol (one screen of output)

  1. State the contradiction in one line, technical form ("improving A vs worsening B") or physical form ("X must be both A and not-A"). If you cannot, stop: there is no contradiction, proceed normally.
  2. Ideal Final Result: what if the function existed with no new code? Name what would have to be true. If reachable, that is candidate 0 and usually the answer.
  3. Reuse inventory: what already does ≥80% of this? path:line or none found.
  4. Prior resolutions + principle prompts:
    PLUGIN_DIR="${CLAUDE_PLUGIN_ROOT:-$(cat "${NAVIGATOR_CONFIG_HOME:-${XDG_CONFIG_HOME:-$HOME/.config}/navigator}/plugin-root" 2>/dev/null)}"
    [ -d "$PLUGIN_DIR/skills" ] || PLUGIN_DIR="$HOME/.claude/plugins/marketplaces/navigator-marketplace"
    python3 "$PLUGIN_DIR/skills/nav-triz/functions/triz_suggest.py" \
      --contradiction "<improving A> vs <worsening B>"
    Prints decisions in this graph that resolved a related tension, then three principle prompts from different separation modes (generic trio if nothing matches). Read reference/PRINCIPLES.md only if a prompt is unclear.
  5. Three candidates, one per prompt, in the table below. Each MUST name what it worsens. A candidate with no downside is a compromise misdescribed, or the IFR.
  6. Recommend one with the reason in one sentence. Max one open question.
  7. After the work, if the resolution was non-obvious, capture it:
    python3 "$PLUGIN_DIR/skills/nav-graph/functions/graph_manager.py" --action add-memory \
      --memory-type decision --summary "<what was chosen>" --concepts "<a,b>" \
      --contradiction "<A vs B>" --separation <time|space|condition|level> \
      --principle "<principle: concrete move>"

Output Template

┌─ TRIZ: <short title> ─────────────────────────────────────────────┐
│ Contradiction  improving <A> vs worsening <B>                       │
│ IFR            <no-new-code version; reachable? yes/no>             │
│ Reuse          <path:line does X> | none found                      │
│ Prior          mem-NNN (<separation>) | none                        │
├───┬────────────┬───────────────────────────┬──────┬────────────────┤
│ # │ mode/prin. │ what changes              │ cost │ worsens        │
├───┼────────────┼───────────────────────────┼──────┼────────────────┤
│ 1 │ time / 10  │ ...                       │ S    │ ...            │
│ 2 │ cond / 15  │ ...                       │ M    │ ...            │
│ 3 │ level / 1  │ ...                       │ L    │ ...            │
├───┴────────────┴───────────────────────────┴──────┴────────────────┤
│ Recommend  #<n> — <one-sentence reason>                            │
└────────────────────────────────────────────────────────────────────┘
Open question (0-1): ...

Rules: three candidates, not five; different separation modes, not three variants of one move; each row's "worsens" cell is mandatory; total output fits one screen.

Hand-offs

  • nav-brief → here when Contradict ≠ none on a substantial task. The brief's Approach cites the recommended candidate.
  • navigator-research agent Phase 0.5 already produces IFR + reuse inventory; when it reports a contradiction, run this protocol in the main session (agents do not invoke skills).
  • Task Mode RESEARCH phase lists this as the divergence step before PLAN.

Predefined Functions

functions/triz_suggest.py --contradiction "<A vs B>" [--graph-path P] [--limit 3] [--format text|json] Deterministic, stdlib, always exits 0. Keyword-ranks the ~20 mapped principles, picks one per separation mode first, pads with prior action / dynamization / segmentation, and lists prior decisions from the graph via query_contradictions.

Related

  • reference/PRINCIPLES.md — separation modes + mapped principles (lazy-load)
  • skills/nav-brief/SKILL.md — Contradict row
  • skills/nav-graph/SKILL.md Step 3B — capturing tagged decisions

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.