agentleFS
chuckplayerchuckplayer/claude-agent-pack/skills/interview-me/SKILL.md

interview-me

Skill8 starsChanged 35 days agoMITView on GitHub

Interview the user about a plan or decision through structured one-at-a-time questioning until reaching consensus. Produces a design brief artifact or hands off when the user signals readiness. Use when an idea is vague and needs shaping before planning or implementation begins. Trigger this when someone says: interview me, help me think through this, I have an idea but I don't know how to approach it, ask me questions about this, what should I be thinking about here. Do NOT use when the task is already well-scoped — use /plan instead.

101 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use constraint-driven-development, idea-refine and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • MrCipherSmithSkill

    constraint-driven-development

    Use when a project needs a persistent, measurable quality bar — starting a project without documented standards, 'set up constraints', or when coverage/perf/a11y thresholds get re-debated per-PR. Produces CONSTRAINTS.md with enforceable numbers, ratchets, and a guard against silent weakening. NOT for throwaway prototypes or a one-off CI setup.

    Also in 16 projects that use this

  • LLl0k0laDSkill

    idea-refine

    Refine ideas through structured divergent and convergent thinking. Use "idea-refine" or "ideate" to trigger.

    Also in 42 projects that use this

  • LLl0k0laDSkill

    using-agent-skills

    Use when starting a session or when you need to discover which skill applies to the current task. This is the meta-skill that governs how all other skills are discovered and invoked.

    Also in 32 projects that use this

  • HK-hubSkill

    observability-and-instrumentation

    Instruments code so production behavior is visible and diagnosable. Use when adding logging, metrics, tracing, or alerting. Use when shipping any feature that runs in production and you need evidence it works. Use when production issues are reported but you can't tell what happened from the available data.

    Also in 32 projects that use this

  • orenluxySkill

    implementation-notes

    Maintain an IMPLEMENTATION_NOTES.md decision log during any multi-step implementation session. Use this in EVERY session that involves writing or modifying code across more than one file, or any session expected to run longer than a few tool calls. Records decisions, deviations from plan, and open questions so nothing is silently guessed.

    Also in 3 projects that use this

  • HK-hubSkill

    deprecation-and-migration

    Manages deprecation and migration. Use when removing old systems, APIs, or features. Use when migrating users from one implementation to another. Use when deciding whether to maintain or sunset existing code.

    Also in 29 projects that use this

  • +6 more from LLl0k0laD

The file

We split this skill into the rules it gives an agent

12 rules. 0 are common, 11 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 87 edited versions we read keep these.

  • 11

    Distinctive

    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. Distinctive:

    Ask exactly one question to understand the type of work:

    line 12
  2. Distinctive:

    Do not ask anything else yet.

    line 16
  3. Distinctive:

    If the work type from step 1 involves code (feature, refactor, architectural decision with code implications), read relevant files before …

    line 20
  4. Distinctive:

    Never ask about a branch that depends on an unresolved upstream decision.

    line 44
  5. Distinctive:

    Stop asking when either condition is met, then route accordingly:

    line 48
  6. Distinctive:

    If the user did NOT signal direct readiness (condition a), write a design brief and present it for approval before writing:

    line 55
  7. Very specific:

    Ask: "Does this capture it correctly?

    line 69
  8. Distinctive:

    Write to the confirmed path after approval.

    line 71
  9. Distinctive:

    After the interview (with or without artifact):

    line 75
  10. Distinctive:

    Passive agreement: If the user keeps saying "yes" without engaging, slow down — "Any concern with that choice, or are you happy with it?" …

    line 103
  11. Distinctive:

    Don't over-interview: If all major branches resolve early, don't invent questions to fill space.

    line 106
  12. Distinctive:

    Don't ask the Stage 0 question on signal (a): When the user says "let's go", do not ask whether they want a spec of record — they just …

    line 107

“Common” is measured: kept by at least 6 in 10 of the 87 edited versions of this file we could read. “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

Interview Me

Shape a vague idea into an actionable brief by settling foundational decisions before branching ones. Silently reads the codebase when work has code implications, and terminates when the user signals readiness or all major branches resolve. Produces an optional design brief that feeds cleanly into /plan or /implement.

1. Open with a single orienting question

Ask exactly one question to understand the type of work:

"What are we working through — a new feature, an architectural decision, a refactor, or something else?"

Do not ask anything else yet.

2. Explore the codebase silently (code-related work only)

If the work type from step 1 involves code (feature, refactor, architectural decision with code implications), read relevant files before forming questions — tech stack, existing patterns, affected modules, related code.

Surface what you found as context, not as questions:

"I can see you're using a repository pattern with EF Core — I'll factor that in."

Skip this step for non-technical work (process decisions, organizational changes, etc.).

3. Ask questions one at a time, in dependency order

Work through decisions in this order:

  1. Goal — What outcome defines success?
  2. Constraints — What can't change? (performance, compatibility, deadlines, compliance)
  3. Approach — How will we get there?
  4. Success criteria — How will we know it worked?
  5. Risks — What could go wrong?
  6. Implementation details — Specifics that depend on the above being settled

For each question:

  • State in one sentence why it matters
  • Offer a recommended answer with brief reasoning
  • Let the user confirm, redirect, or skip

Never ask about a branch that depends on an unresolved upstream decision. Hold it until the dependency resolves.

4. Detect the termination signal

Stop asking when either condition is met, then route accordingly:

  • (a) User signals readiness — "let's go", "implement it", "that's enough", "I'm ready" → skip step 5, proceed directly to step 6
  • (b) All branches resolved — goal, constraints, approach, success criteria, and key risks all have answers → proceed to step 5 to produce the brief

5. Produce the artifact (optional)

If the user did NOT signal direct readiness (condition a), write a design brief and present it for approval before writing:

## Design Brief: <title>

**Goal:** <what success looks like>
**Constraints:** <what can't change>
**Approach:** <chosen strategy and why>
**Key decisions:**
- <decision>: <rationale>
**Known risks:** <flagged but accepted>
**Open questions:** <anything deferred>

Ask: "Does this capture it correctly? Any changes before I write it? I'll write to docs/<slugified-title>-brief.md — does that path work, or would you prefer somewhere else?"

Write to the confirmed path after approval.

6. Hand off

After the interview (with or without artifact):

  • Approach still ambiguous → offer /plan to decompose with tech-lead
  • Plan clear enough to build → offer /implement with the brief as context
  • User wants to iterate → loop back to step 3

Stage 0 artifact — gated on step 4's signals, and no others

  • Signal (b), all branches resolved → ask, once:

    "Do you want a spec of record for this — requirements captured with source locators and a field inventory — or just build it?"

    On yes, route to /spec-intake with the written brief path as the input. On no, continue with the hand-off above and say nothing further about it — asking once is the whole mechanism. Name the artifacts, not the pipeline, because the artifacts are what the user is choosing between.

  • Signal (a), the user said "let's go" / "implement it" → do not ask. Note it once, in one line — "No Stage 0 artifact for this, so /verify-spec will have nothing to check it against later." — then proceed to /implement.

The mechanism differs from the other two branches on purpose. /plan versus /implement is inferable from interview state, whereas whether this work is a tracked deliverable is an organizational fact that no amount of interview content answers — so it has to be asked, and asked only where asking is still welcome.

Gotchas

  • Passive agreement: If the user keeps saying "yes" without engaging, slow down — "Any concern with that choice, or are you happy with it?" Uncritical agreement is not consensus.
  • "I don't care" is not a decision: Offer a concrete default and confirm — "I'd default to X, does that work?" Don't leave the branch open.
  • Don't invoke tech-lead mid-interview: Discovery must complete before handoff. Calling tech-lead while questions remain open produces a half-baked plan.
  • Don't over-interview: If all major branches resolve early, don't invent questions to fill space. Terminate and hand off.
  • Don't ask the Stage 0 question on signal (a): When the user says "let's go", do not ask whether they want a spec of record — they just declined ceremony, and re-imposing it contradicts the gotcha directly above. Emit the one-line note and proceed.

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.