agentsclimarketplace

To spec

Skill shikanime-labs/skills/skills/workflow/to-spec

A curated catalog of self-improved agent skills for Hermes and compatible agentsFrom the repository description

Install
npx -y skills add shikanime-labs/skills --skill to-spec

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

3 things to look at

  • skips confirmationTells the agent to proceed without asking first, 1 time: "Do **not** interview the user — synthesize what you already discussed.".
  • 2 stars2 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.
  • runs commandsInstructs the agent to run 2 commands, including `bootstrap` and 1 more.

What its file declares

Copied from the file, not written here

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

3.4 KB, 715 tokens by cl100k_base, as published. Nobody here has run it

<!-- markdownlint-disable MD013 MD033 -->

To Spec

Take the current conversation and codebase understanding and produce a spec (a PRD), then publish it as a GitHub issue. This is the validation gate of the workflow: a feature exists as a reviewed spec issue before any code is written.

Do not interview the user — synthesize what you already discussed. The issue tracker and label vocabulary come from docs/agents/issue-tracker.md; run bootstrap if it's missing. Use the glossary in docs/agents/domain.md throughout, and respect ADRs in the area you touch.

Process

  1. Explore the repo to understand current state if you haven't already. Stay on the minimalism ladder (docs/agents/workflow.md): name the minimum seams and the minimum interface the feature needs — never speculative scaffolding or "for later" flexibility.
  2. Update the strategic docs so the feature is maintainable by humans and other agents. Sharpen any new terminology into CONTEXT.md and, if the feature commits to a non-obvious approach, record it as an ADR in docs/adr/ (per docs/agents/domain.md). The spec issue references these.
  3. Sketch the seams at which you'll test the feature. Prefer existing seams to new ones; use the highest seam possible; fewer is better (ideal: one). Name the feature-driven test(s) and the property-based stability invariant(s) you expect implement to satisfy. Confirm the seams match the user's expectations.
  4. Write the spec using the template, then publish it as an issue. This issue is the meta issue for the work (strategic, like an epic) — label it ready-for-human so a human approves direction before any code. Set its milestone (the target release), assignee, project, type, and priority at creation; keep the body to relevant content only (per docs/agents/issue-tracker.md). Hand back the issue URL; to-tickets parents the tactical tickets to it.
<spec-template>

Problem Statement

The problem the user faces, from the user's perspective.

Solution

The solution, from the user's perspective.

User Stories

A long, numbered list, each: As an <actor>, I want <feature>, so that <benefit>. Cover all aspects of the feature — be extensive.

Implementation Decisions

Modules built/modified, their interfaces, technical clarifications, architectural decisions, schema changes, API contracts, specific interactions. Do NOT include file paths or code snippets (they go stale fast) — the exception is a snippet from a prototype that encodes a decision more precisely than prose.

</spec-template>

Publishing

Create the issue via the GitHub MCP (GitHub_issue_write) or gh issue create, per docs/agents/issue-tracker.md, setting labels, milestone, assignee, and project. Title it in the domain's vocabulary. Hand back the issue URL.

Boundaries

Produces one spec issue. Does not break it into tickets (that's to-tickets) and does not implement it (that's implement).

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 325,949. 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.