agentleFS
gleicongleicon/devskills/skills/ds-explore/SKILL.md

ds-explore

Skill20 starsChanged 46 days agoMITView on GitHub

Suggest candidate approaches to a problem — research and lay out options, but never decide.

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:

    Read the context so options respect reality:

    line 17
  2. Very specific:

    Write the artifact to EXPLORE.md in the current directory.

    line 25
  3. Very specific:

    Do not edit code or anything under .project/.

    line 30

≈ “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

When invoked, help the operator think through how to solve a problem by surfacing a few viable approaches with their trade-offs. You suggest; you do not decide and you do not implement. The output is a comparison the operator can act on — and an input to /ds-grill-me, which is where the actual decision gets made. (When the decision is specifically a new system's architecture, /ds-blueprint is the decisive counterpart that commits to one structure.)

Arguments

  • A problem or question to explore. If none is given, ask once, then proceed.
  • --web — opt-in: do bounded web research (a few high-signal sources, each cited). Off by default; the command works from the project context and your own knowledge.

Process

  1. Establish the problem. If the user gave none, ask once.
  2. Read the context so options respect reality:
    • .project/state.md if present — honor what # settled and # hazards record, don't re-litigate them.
    • the relevant code.
  3. Gather information:
    • With --web: research, bounded — a handful of sources, cite each, no open-ended crawl.
    • Without --web: work from the context and your knowledge. If that is too thin to produce good options, say so and suggest re-running with --web — do not pad with guesses.
  4. Lay out 2–4 candidate approaches. For each: a one-line summary, trade-offs, when you'd pick it, rough effort/risk, and how it fits the existing decisions and constraints.
  5. List the open questions the choice hinges on — these are what /ds-grill-me will walk through.
  6. Write the artifact to EXPLORE.md in the current directory; return the path.

Rules

  • Suggest, never decide. If you have a lean, state it and mark it explicitly as the operator's call.
  • Do not edit code or anything under .project/. This command only reads, and writes its own artifact.
  • Build options on top of what is already settled — do not reopen a call state.md records.
  • Stay bounded: 2–4 options, not an exhaustive survey. Cite any web sources.
  • EXPLORE.md is a scratchpad — overwrite it. It never goes in .project/, which holds session state, not work products.

Output

The candidate approaches and open questions inline, plus the artifact path. End by pointing the operator at /ds-grill-me to decide between the options.

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.