fix
/fix
Write a current-feature.md spec for an ad-hoc bug fix or small unplanned change, then stop before implementation. Use for /fix, confirmed bugs, or small changes that do not belong in the build plan.
- Author
- A7-studio ↗
- Repository
- jobmigo ↗
- Licence
- MIT
- Last repository change
- 12.11.2026
Assessment and security
- Files in this package
- 1
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 12.11.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
Write a current-feature.md spec for an ad-hoc bug fix or small unplanned change, then stop before implementation. Use for /fix, confirmed bugs, or small changes that do not belong in the build plan.
Original author description
Write a current-feature.md spec for an ad-hoc bug fix or small unplanned change, then stop before implementation. Use for /fix, confirmed bugs, or small changes that do not belong in the build plan.
How it works
- Input
- Step 1 - write the fix spec
- Rules
- Formatting
SKILL.md preview
---
name: fix
description: Write a current-feature.md spec for an ad-hoc bug fix or small unplanned change, then stop before implementation. Use for /fix, confirmed bugs, or small changes that do not belong in the build plan.
---
# fix - document an ad-hoc fix, then build it like anything else
**Context reuse:** Reuse any required file already loaded in project instructions or the current session. Read it again only if absent, changed, or exact current bytes or line references are needed.
**First action:** Before project inspection, preflight, or any other tool call,
publish `running` to `blueprint/.state/run.json` when `blueprint/.state/`
exists. From the project root, run the first helper that exists,
`node .agents/skills/blueprint-doctor/scripts/run-state.mjs` or
`node .claude/skills/blueprint-doctor/scripts/run-state.mjs`, with `start --command
<skill> --summary "<truthful summary>" --boundary
<read-only|reviewed|local-only>`. Use `update` at milestones, with `--status
blocked --resume <command>` for a blocker, and `finish --status ready|completed`
at the end; `--help` lists every option and `blueprint/workflow.md` has the full
activity contract. A missing or failing helper is a warning, not a blocker.
Where this sits in the workflow:
/fix -> /implement -> /complete -> back to your features
(spec (build it, (log to blueprint/history/fixes/
the fix) reviewed) + merge)
A fix is a bug or small change that isn't a planned build-plan feature. It runs
through the same loop as a feature (build with review gates, iterate, then merge);
it just starts here instead of `/feature`, and is logged separately.
## Input
A description of the bug or change, for example `/fix "password reset email never
sends"`. If the user just reported the problem in chat, use that.
The input may also be a finding ID from `blueprint/context/findings.md`, alone
or with a description, for example `/fix F-03`. Pull the problem statement from
that ledger entry. Use this form only between work items, when
`current-feature.md` is the reset stub: this skill overwrites that file, so
while a spec is active, repair its findings through `/implement` instead.
Never replace another worker's active spec. Parallel work uses one clone or Git
worktree per work item, with a separate branch in each checkout.
## Step 1 - write the fix spec
Pull context from `blueprint/context/project-overview.md` and `blueprint/context/coding-standards.md`,
then write a short spec to `blueprint/context/current-feature.md` (this file holds whatever
is being built now, feature or fix). Its first heading must be exactly
`# Fix: <title>`, and it must retain the `**Type:** Fix` contract below. Keep it
lighter than a feature spec:
- **Title** - the bug or change in a few words.
- **Type:** Fix (so `/complete` logs it to `blueprint/history/fixes/`, not `blueprint/history/features/`).
- **Status:** not started - `/implement` updates this durable workflow state as
work and verification progress.
- **Branch:** the full fix branch from the configured prefix plus the fix title
in lowercase kebab-case.
- **Fixes:** `<finding id>` - only when the fix targets a ledger finding. The
stamp makes the repair traceable: `/implement` marks that finding `fixed`
when the repairing step lands, and `/audit` re-reviews it before it closes.
- **The problem** - what's wrong or what needs to change, and where.
- **The fix** - the root-cause repair, and anything it must not break. Do not add
an abstraction, dependency, service, configuration surface, compatibility
layer, or security mechanism unless the defect is in that layer or an
established repository requirement needs it now. Adding a missing
authorization, ownership, validation, escaping, or redaction check is a
root-cause repair, not new machinery.
- **Build steps** - usually one small step; split only if the diff would be too
big to read. Each ends with an observable "done when".
- **Verify** - how to confirm it's fixed (what to click or test).
Then stop. Tell the user to review the fix spec, then run `/implement` to build it.
## Rules
- A fix is not a build-plan item; don't add it to `build-plan.md`.
- Keep it small. If it's really a new feature, use `/feature` and the build plan
instead.
- Same conventions as everything else (`blueprint/context/coding-standards.md`).
## Formatting
Format the output to match the project's conventions in
`blueprint/context/ai-interaction.md`: concise, scannable markdown, with lists for
enumerations and tables for matrices rather than dense paragraphs.
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · A7-studio/jobmigo | 100 | MIT |
| swingerman/engineer | 100 | MIT |
| Hainrixz/claude-seo-ai | 100 | MIT |
| vanducng/skills | 100 | MIT |
| RemiPelloux/kill-my-idea | 100 | MIT |
| tony/skills | 100 | MIT |
| applelog/nondev-kit | 100 | MIT |
| koda-ai-studio/agent-bom | 100 | Apache-2.0 |
| bendrucker/claude | 100 | MIT |
Database snapshot: 13.11.2026 · package version 0a7ed925f9eb (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 100scope-confirmed-changes
Use when fixing a bug, a regression, or a confirmed review finding. Also use it when the user says "fix this bug", "it is broken", or "keep the change small". Reproduces the issue before any fix, with a test, a real command, or the real interface. After the fix, the same… - 100writing-feature-prompts
Write a ready-to-run implementation prompt for a feature request or a bug fix in the current repository and save it outside the repo. A feature prompt runs whatever design, requirements, documentation and incremental test-first implementation phases the repo's own rules call… - 100bug-fix
Fix a Device bug the project's way - reproduce it, write a small failing test, find the root cause, fix it, prove the test now passes. Use for any bug report or broken feature; escalates to a spec when the cause or fix is not obvious. - 100rollback
Plan a safe, history-preserving reversal of a completed feature from its exact work commit, including later-dependency risk, then stop before implementation. Use for /rollback or requests to undo, remove, or reverse a completed feature. Not for a bug in a live feature (/fix) or diagnosis (/debug). - 100planning-flow-quick
Use when the user wants a quick plan for a small coding change — "quick plan", "plan this small fix", "/planning-flow-quick <ask>" — such as a renamed title, a corrected message, a one-function bug fix, or a small option, where the change still needs to follow the repository's… - 90feature
Implement a feature from a spec (docs/specs/NNN_<name>.md), a spec slice, or a GitHub issue: write the code and tests, run the checks, and stop with a working tree to review. --commit also commits; --pr also opens the PR and runs its review. Use for 'implement the auth spec',…