agentleFS
Sign inStart free
DeiblerDeibler/edmund-harness/skills/teams/compare-options/SKILL.md

compare-options

Skill24 starsChanged 2 months agoMITView on GitHub

Parallel-investigator team for side-by-side comparisons. One agent per option, investigating in isolation so their findings don't contaminate each other. Use when the user asks "which of X, Y, Z should I pick?" and each deserves a real look.

2 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

3 rules. About 0 are common, 1 is distinctive, and 2 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.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Don't use for:

    line 17
  2. Very specific:

    If criteria differ per option, you're not really comparing — split into separate spawn_agent calls.

    line 44
  3. Very specific:

    If the user wanted raw data, offer to send the full markdown files via send_attachment.

    line 53

≈ “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

compare-options

N agents, each investigating one option with the same set of questions, so the comparison is apples-to-apples.

When to use

  • "Should I use Framework A or Framework B?"
  • "Which of these 3 places has the best [criterion]?"
  • "Compare these 4 iMessage servers."
  • Any question where the user named the options and wants them evaluated side-by-side.

Don't use for:

  • Open-ended discovery ("what frameworks exist for X?") → that's research-trio.
  • One-option deep-dives → single spawn_agent is lighter.

The pattern

Spawn one agent per option, all with the SAME evaluation criteria. Each writes its findings to {shared}/<option-slug>.md. No cross-talk; each agent doesn't know the others exist.

spawn_team(members: [
  {role: "option-a", task: "Evaluate {OPTION_A} against these criteria: <criteria list>. Write findings to {shared}/option-a.md with sections per criterion. Final text: one-paragraph verdict."},
  {role: "option-b", task: "Evaluate {OPTION_B} against these criteria: <criteria list>. Write findings to {shared}/option-b.md with sections per criterion. Final text: one-paragraph verdict."},
  {role: "option-c", task: "Evaluate {OPTION_C} against these criteria: <criteria list>. Write findings to {shared}/option-c.md with sections per criterion. Final text: one-paragraph verdict."}
])

Criteria matter

Agents investigate what you tell them to. Write the criteria list explicitly in every member's task — same list for each:

  • "Install ease (1-10)"
  • "Active maintenance (last 6 months)"
  • "macOS 15 Sequoia compatibility"
  • "License"
  • "Known issues"
  • Whatever the user cares about.

If criteria differ per option, you're not really comparing — split into separate spawn_agent calls.

After completion

read_team_results(team_id) returns each agent's verdict. Synthesize:

  1. One-sentence recommendation ("go with X because Y").
  2. Short side-by-side: one bubble per option, 2-3 words on each criterion.
  3. The main tradeoff between the top 2.
  4. If the user wanted raw data, offer to send the full markdown files via send_attachment.

Tradeoffs

  • N API calls. For 3 options this is ~$0.10-0.30, 30-60s wall clock. For 5+ options, think about whether this actually helps vs a single well-prompted research-trio.
  • Strongest when criteria are fixed and options are named. Weak when the user is still deciding what criteria matter.

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.

Check your setup, free