agentsclimarketplace

Mecris plan

Skill kingdonb/mecris/.claude/skills/mecris-plan

Write a minimal spec as a GitHub issue before acting. Takes the recommended action from mecris-orient and records intent, rationale, and validation criteria. This is the red in red-green-archive. Trigger with /mecris-planFrom its SKILL.md

Install
npx -y skills add kingdonb/mecris --skill mecris-plan

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

One thing to look at

  • 4 stars4 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

3.4 KB, as published. Nobody here has run it

mecris-plan

Spec-writing skill for the Mecris accountability system. Before any code is touched or action is taken, write a three-line GitHub issue on yebyen/mecris recording what will be done, why, and how to know it worked. Consumes the orient report. Produces an auditable spec.

When to Activate

  • After /mecris-orient has produced a situation report with a recommended action
  • Before writing any code, making any commit, or triggering any workflow
  • When the user says "plan", "write the spec", "what's the plan", "/mecris-plan"

Slash Command

/mecris-plan

Runs the full planning workflow:

  1. Take the recommended action from the current orient report (must have run /mecris-orient first)
  2. Draft a three-part spec:
    • Intent: one sentence — what will be done
    • Because: one sentence — what in the orient report motivates this
    • Validation: one sentence — how to confirm it worked
  3. Open a GitHub issue on yebyen/mecris with title [plan] {intent} and body containing all three parts
  4. Record the issue number — archive will close it when work is complete

Usage: Type /mecris-plan after orienting. The spec issue will be opened before any action proceeds.

Spec Format

Issue title: [plan] {one-line description of action}

Issue body:

**Intent**: {what will be done — specific and concrete}

**Because**: {what orient showed that motivates this — cite the source: issue number, commit, or NEXT_SESSION.md item}

**Validation**: {observable outcome that confirms success — test pass, issue closed, datapoint in Beeminder, etc.}

Examples

Sync from upstream

[plan] Sync yebyen/mecris from kingdonb/mecris main (18 commits behind)

**Intent**: Fetch and merge kingdonb/mecris main into yebyen/mecris.

**Because**: Orient showed yebyen/mecris is 18 commits behind upstream with no merge conflicts anticipated.

**Validation**: `git rev-list --count HEAD..upstream/main` returns 0.

Fix a bug

[plan] Fix multiplier lever SQL query to respect user_id column

**Intent**: Update the SQL in the Rust service to use the composite primary key (user_id, language_name).

**Because**: NEXT_SESSION.md lists multiplier sync as unverified; issue #122 reports persistence race condition.

**Validation**: `SELECT pump_multiplier FROM language_stats WHERE user_id = ?` returns the value set in the Android app.

Trigger pr-test

[plan] Run pr-test for kingdonb/mecris#142

**Intent**: Dispatch the pr-test workflow against PR #142 and post results as a comment.

**Because**: Orient found issue tagged needs-test referencing PR #142.

**Validation**: Comment posted on kingdonb/mecris#142 with ✅ or ❌ for both Python and Android tests.

Notes

  • This skill opens exactly one issue per invocation — no more
  • The issue stays open until /mecris-archive closes it
  • If orient did not run first, run /mecris-orient before proceeding
  • In bot context: the plan issue is the only thing an outside observer can see before the work starts — make it legible to a human
  • Keep specs small: if the intent requires more than one sentence, the action is too large — split it

Keep looking

Skills are one crate of 627,801. 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.