agentsclimarketplace

Compact shaping

Skill clsaa/littlepowers/skills/compact-shaping

Proportional planning, recovery, and engineering-discipline protocol for Codex, Claude Code, Qoder, and OpenCode.

Install
npx -y skills add clsaa/littlepowers --skill compact-shaping

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

  • 22 days oldThe repository was created 22 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
  • 0 stars0 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.

What its author says it does

Copied from the file, not written here

Create one concise shape brief. Use only when routed by using-littlepowers or active state phase=shape.

SKILL.md

3.0 KB, 589 tokens by cl100k_base, as published. Nobody here has run it

Compact shaping

Resolve the few decisions that matter without producing four separate documents or narrowing the approved outcome.

Read the active ledger, highest-authority parent acceptance sources, approved baselines, and relevant repository evidence. Ask only when a missing answer materially changes behavior, cost, risk, or scope. Do not reopen settled choices.

For an existing workflow, keep the artifact root already resolved by using-littlepowers. When resolving a new workflow, use a non-default root only when the latest user request or a current repository instruction explicitly names it for new workflow artifacts. Existing directories, backlinks, and historical or tool-branded paths do not qualify by themselves. Otherwise use docs/littlepowers/shapes/YYYY-MM-DD-<slug>.md. Include:

  • measurable outcome and non-goals;
  • constraints and low-risk assumptions;
  • requirements and acceptance checks;
  • selected approach and decision rationale;
  • affected files or components;
  • ordered execution steps and validation commands.

Include the inherited complete outcome and Added / Changed / Deferred / Removed, or No scope delta. Do not create product or technical slices; ordered steps are implementation order only. Highlight and obtain explicit approval for any non-empty scope delta.

Read ../../references/outcome-lock.md. Include one Outcome Contract and one Outcome Plan Map in the shape. Give each complete observable outcome a stable OUT-### ID and map every active ID to a task and named evidence. The shape is one artifact, not a reason to add a specification, design, or separate plan.

Keep the brief proportional. If shaping exposes material unresolved architecture, security, migration, cross-system, irreversible-state, or costly-rollback choices, switch to the full route before implementation.

Checkpoint with the current workflow ID and revision:

<python> <state-cli> checkpoint \
  --workflow <workflow-id> --expect-revision <revision> \
  --phase shape \
  --artifact shape=<artifact-path> \
  --completed "compact shape" \
  --progress "Compact shape complete; execution is next" \
  --next-action "Review, bind, and validate the compact shape"

Use the returned revision, then present the shape brief for review and stop: summarize the selected approach and execution steps, name the artifact path, and name executing-plans as the next phase. After approval, run bind-contract --artifact <artifact-path> --approval-kind review-gate with --approve-scope-delta only for a distinctly approved delta, then validate-plan --artifact <artifact-path>. For explicit unattended end-to-end authorization use the matching approval kind. After both commands succeed, checkpoint --phase execute and invoke executing-plans. When the user asks for corrections, revise this artifact, checkpoint again, and present it again instead of advancing.

What ships with it: 1 file

191 B alongside SKILL.md

agents/

Keep looking

Skills are one crate of 326,984. 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.