build-rescue
Skill1 starsChanged 29 days agoMITView on GitHub
Stage 4 of the build skillset, on demand and repeatable. Writes <FEAT>/40-rescue.md — scans code, CONTEXT.md and decisions/ for entropy and concept/code drift, applies deepening refactors one at a time within budget with revert on regression, and reports the concept changes needed. Called by the build Director after synthesis has produced at least one slice.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- voyager67/AgentFieldkittheir own version, first seen Nov 8, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- voyager67/AgentFieldkit★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository6 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- build-conceptSkill
- build-intakeSkill
- buildSkill
- build-synthesisSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 28
Apply, one at a time: Run the full test suite before.
- Distinctive: line 29
Reconcile: Every kept refactor that changes a signature, module boundary or term is a concept change: list the exact sections.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
rescue — Stage 4
You are the maintenance crew: you make the codebase simpler without changing what it does, and you report where concept and code have drifted. Read ../../RULES.md first (§10 binds you); the process is ../../MANIFEST.md Phase 10 and "Concept ↔ Code Synchronization".
Input: <FEAT>/20-concept.md, <FEAT>/30-synthesis.md, CONTEXT.md, decisions/, test_tooling.md, code_root. Output: <FEAT>/40-rescue.md (a new .vN per round) from templates/40-rescue.md; refactored code under code_root. Budget: budget_min.rescue; budget_refactors = 5 with git, 2 without.
Lineage: MANIFEST Phase 10; maquette-harden (one commit per fix, revert on regression).
Opening
Two sentences: a maintenance round — find what got tangled, untangle the most valuable pieces one at a time, and say what the concept must learn from it.
Sequence
- Scan. Read the concept's traceability matrix against the code: modules that exist but are not in §11, behaviour with no requirement, requirements whose module changed shape. List drift with IDs. Then look for deepening opportunities: interfaces that could be simpler, duplicated logic, names that left
CONTEXT.md, dead code, modules doing two things. Rank by expected gain × low risk. - Propose. Show the ranked table and ask: "which ones, in this order?" — one question.
- Apply, one at a time. Run the full test suite before; apply the refactor; run it after. Green → keep (with git: commit, RULES §6). Any regression → revert immediately, record the reason, move on. Stop at the refactor budget or the time budget and say which.
- Reconcile. Every kept refactor that changes a signature, module boundary or term is a concept change: list the exact sections; the Director schedules a concept redo (version bump). Add ADRs for decisions taken here. Update
test_tooling.mdif a tool changed. - Write
40-rescue.mdand hand back per RULES §7 with: kept / reverted counts, drift items, concept sections to update.
Anti-patterns
| Don't | Do instead |
| Refactor and add features in one go | Behaviour-preserving only; features are slices |
| Batch several refactors into one commit | One each; revert is cheap only when isolated |
| Fix drift by editing the concept yourself | Report the sections; concept redo is the concept stage's job |
| Keep a refactor with a red test "to fix later" | Revert; note it as an opportunity for the next round |
Rename without updating CONTEXT.md | The dictionary moves with the code |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.