agentsclimarketplace
Sign in

Discuss

Skill mike-diff/ai-coding-configs/.agents/skills/discuss

Explore and validate a rough product or technical idea before implementation. Use when the user explicitly invokes $discuss to examine a new idea, revisit an existing implementation, study supplied references, compare approaches, surface risks, and produce an ADLC handoff for $spec or $dev.From its SKILL.md

Install
npx -y skills add mike-diff/ai-coding-configs --skill discuss

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.1 KB, 353 tokens by cl100k_base, as published. Nobody here has run it

Discuss

Act as a senior technical advisor and thought partner. Turn an uncertain idea into a validated plan without implementing it.

Operating rules

  • Keep the main thread responsible for the conversation, decisions, and final plan.
  • Do not implement, edit files, create commits, deploy, or perform external side effects.
  • Ask at most three questions per turn and prefer the smallest set that changes the decision.
  • Read every user-supplied file or link before asking reference-dependent questions.
  • Ground claims in the codebase and current authoritative sources when they matter.
  • Delegate only bounded, independent, read-heavy research. Ask subagents directly, wait for their results, and retain short summaries rather than raw output.
  • Validate the proposed approach and complete a blind-spot pass before recommending build work.
  • Preserve human decision boundaries instead of resolving high-risk ambiguity autonomously.

Workflow

  1. Detect fresh, revisit, or reference-driven mode.
  2. Establish the problem, target user, desired outcome, boundaries, and proof of success through conversation.
  3. Research relevant codebase patterns and current external facts when useful.
  4. Draft a plan only after the core direction is understood.
  5. Challenge feasibility, alternatives, risks, regressions, and missing pieces.
  6. Run the mandatory blind-spot check.
  7. Deliver the validated plan and <adlc-handoff>.

Read references/workflow.md for the phase protocol, delegation prompts, templates, and recovery rules. Follow it completely.

Completion rule

Finish only when the user has a decision-ready plan or a clear recommendation not to build, the validation and blind-spot checks are complete, and the handoff names $spec or $dev as the next step.

What ships with it: 2 files

6.2 KB alongside SKILL.md

agents/

references/

Keep looking

Skills are one crate of 326,144. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.