agentleFS
JlosevJlosev/agent-skills/skills/tool-market-scout/SKILL.md

tool-market-scout

Skill1 starsChanged 17 days agoMITView on GitHub

JTBD brief to community scan to scoring to verdict Buy / Build / Hybrid / Defer. Triggers tool-market-scout, buy vs build, tool market scan. Delegates scan to tool-scout.

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 9 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If the user named a package and only wants docs/API → stop, wrong skill.

    line 37
  2. Distinctive:

    If trigger = product to replace → Step 1a, not alternatives search.

    line 38
  3. Very specific:

    Do not: run your own WebSearch/multi-source scan.

    line 51
  4. Very specific:

    If tool-scout is not installed → ask once: npx skills add LpcPaul/tool-scout-skill.

    line 56
  5. Distinctive:

    If stake = no → shortlist only.

    line 70
  6. Very specific:

    If the user wants a note: Research/Tooling/{topic-slug}.md from references/output-template.md (or path they name).

    line 74
  7. Distinctive:

    No scan without confirmed JTBD brief

    line 78
  8. Distinctive:

    No primary «alternatives to {brand}»

    line 79
  9. Distinctive:

    No reimplementing tool-scout scan

    line 80
  10. Distinctive:

    No verdict without scoring when buy-or-build stake

    line 81
  11. Distinctive:

    No auto-Buy; verdict waits for ack

    line 82
  12. Distinctive:

    When stake: scoring with evidence + verdict with ack

    line 88

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
<!-- Scan: https://github.com/LpcPaul/tool-scout-skill (LpcPaul) --><!-- Verdict rubric: https://mm.dreamineering.com/playbook/applications/select-tech/ (Dreamineering) -->

Tool Market Scout

JTBD brief → community scan (tool-scout) → scoring → verdict Buy / Build / Hybrid / Defer.

{SKILL_DIR} = directory of this SKILL.md.

Scope

In: choosing SaaS, desktop/mobile, OSS lib, MCP/skill; constraints (platform, budget, OSS-only, data sovereignty)
Out: docs lookup for an already named package, product strategy for a whole market, customer interviews, internal service TDR
Fallback: if candidates <2 → Build or widen JTBD

Gotchas

  • JTBD brief + ack before scan – references/jtbd-decompose.md
  • No primary «alternatives to {brand}» – capability queries from functional jobs
  • Do not reimplement scan – delegate to tool-scout (Step 2)
  • Scoring + verdict – only when ≥2 candidates and buy-or-build stake
  • No auto-Buy – verdict waits for ack
  • Stars ≠ quality – references/scoring-rubric.md

Algorithm

Step 0: Intake + scope gate

Collect trigger, constraints, must-have capabilities, buy-or-build stake (yes/no).
If the user named a package and only wants docs/API → stop, wrong skill.
If trigger = product to replace → Step 1a, not alternatives search.

Step 1a: Native audit (if a product is named)

Official docs/features → functional jobs, not marketing. Format – references/jtbd-decompose.md.

Step 1b: JTBD brief (gate)

Job statement, functional jobs, constraints, out of scope, buy-or-build stake.
STOP. Show the user. Wait for «ok» / edits. No ack – no scan.

Step 2: Community scan (tool-scout)

Do not run your own WebSearch/multi-source scan. Invoke / run the installed tool-scout skill with capability queries from the brief.

  • Query = capability-first from functional jobs (references/search-sources.md)
  • Supplementary {brand} alternatives – max 1, after brief only
  • CLI, flags, and scripts for tool-scout live in LpcPaul/tool-scout-skill only; do not duplicate here
  • If tool-scout is not installed → ask once: npx skills add LpcPaul/tool-scout-skill; then rerun via the skill. Fallback: minimal capability search (no full scan reimplementation)

Normalize to candidate table (name, type, OSS/paid, source URL, covers jobs).

Step 3: Score (conditional)

Only if ≥2 candidates and buy-or-build stake = yes.
Rubric – references/scoring-rubric.md. ≥3 candidates + Build yourself row. Evidence required per score.

Step 4: Verdict (gate)

Only if Step 3 ran. Buy | Build | Hybrid | Defer; top pick; risks; next step; how to start.
STOP. Show scoring + verdict. Wait for ack. No auto-Buy.

If stake = no → shortlist only; offer verdict on request.

Step 5: Writeback (optional)

If the user wants a note: Research/Tooling/{topic-slug}.md from references/output-template.md (or path they name). Skip if chat-only.

Hard Stop Rules

  • No scan without confirmed JTBD brief
  • No primary «alternatives to {brand}»
  • No reimplementing tool-scout scan
  • No verdict without scoring when buy-or-build stake
  • No auto-Buy; verdict waits for ack

Definition of Done

  • JTBD brief confirmed
  • Scan via tool-scout (or install prompt + fallback)
  • When stake: scoring with evidence + verdict with ack
  • URLs are markdown links to primary sources

Check commands

SKILL_DIR="<path-to-this-skill>"
ls "$SKILL_DIR/references"/*.md

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.