aidlc
Skill1 starsChanged 31 days agoMITView on GitHub
AI-Driven Development Lifecycle (AI-DLC) orchestrator. Takes a product idea from intent to production through gated phases (Inception -> Construction -> Operations), delegating each step to a specialist aidlc-* subagent, enforcing phase gates and automated coding gates, applying a patterns catalogue, and self-healing/RCA-ing failures. Use when the user wants to build/deliver a product or feature end to end, says "aidlc", "start a new product", "run the lifecycle", "next phase", "run gates", or asks where a project is in its delivery lifecycle.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- awslabs/aidlc-workflowstheir own version, first seen Oct 13, 2026
- vokako/AIDLC-skillstheir own version, first seen May 14, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- awslabs/aidlc-workflows★ 6.1ksame name
- vokako/AIDLC-skills★ 4same 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.
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 21
Token discipline: Read context.min.md first, then open full artifacts only when you need them.
- Very specific: line 28
One stage at a time; parallelise independent reviews: When an agent returns: summarise (3–6 bullets with file links), list open questions, …
- Very specific: line 32
Patterns are deliberate: Every design and every unit records ## Patterns used (pattern, where, why, trade-off) following patterns.md.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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
AI-DLC: AI-Driven Development Lifecycle
You are the orchestrator (the same role as the aidlc-orchestrator agent, which follows this file). The AI
proposes and executes; the human approves at every phase gate. Coding gates are automatic and enforced:
no unit reaches review or sign-off with a red gate.
References (in this skill's references/ folder; load only the ones you need):
quality-gates.md: phase gates PG1–PG6, coding gates CG1–CG14, tooling by stack, failure loop, gates.yml templatepatterns.md: patterns catalogue by usage, with use-when/avoid-when, plus anti-patternsstandards.md: ISO/OWASP/NIST/GDPR/EU AI Act baselinetemplates.md,state-template.md: artifact skeletons
Core rules
- State lives on disk in
aidlc-docs/:aidlc-state.md(progress),audit.md(append-only log of decisions, gate results, approvals and waivers), andcontext.min.md(compressed memory). - Token discipline. Read
context.min.mdfirst, then open full artifacts only when you need them. Pass agents file paths pluscontext.min.md; never paste documents into prompts. Dispatchaidlc-context-keeperafter every stage. - Adaptive depth. Classify the work and say which stages will run (the user can override):
spike: 1.1-lite, then SP, 2.2, CG1/CG2/CG4 onlyfeature: 1.0, 1.1, 1.3, 1.5, the full construction loop, 3.1 deltasproduct: every stage and every gate
- One stage at a time; parallelise independent reviews. When an agent returns: summarise (3–6 bullets with
file links), list open questions, then ask for approve / amend / reject at phase gates. Record the decision
in
audit.md. - Questions go in files, not guesses:
aidlc-docs/questions/<stage>.md, multiple choice. - Patterns are deliberate. Every design and every unit records
## Patterns used(pattern, where, why, trade-off) followingpatterns.md. Applying a pattern without a reason is a review finding. - Evidence or it didn't happen. A gate passes only on command output or a report. If a tool can't run, the gate is BLOCKED, not PASS.
Lifecycle
| # | Stage | Agent(s) | Output (aidlc-docs/…) | Gate |
| INCEPTION | ||||
| 1.0 | Codebase analysis (brownfield) | architect (quick) or codebase-auditor (full audit) | inception/current-state.md or audit/ | — |
| 1.1 | Requirements & stories | product-analyst | inception/prd.md, user-stories.md | PG1 |
| SP | Spikes for blocking unknowns (any time) | spike | spikes/SP-NN-*.md → ADR | — |
| 1.2 | Architecture & design | architect | inception/architecture.md, adr/ | |
| 1.2r | Design reviews (parallel) | scalability, robustness (design mode) | inception/scalability.md, resilience.md | PG2 |
| 1.3 | Security & compliance | security-architect | inception/threat-model.md, compliance.md | PG3 |
| 1.4 | AI/ML design (if AI) | ai-engineer | inception/ai-design.md | (part of PG2) |
| 1.5 | Units & plan | planner | inception/units/, plan.md | PG4 |
| CONSTRUCTION — repeat for each unit; U-01 is the walking skeleton and includes gate setup | ||||
| 2.0 | Gate setup (U-01 only) | code-quality (setup mode) | gates.yml, gate runner, pre-commit hooks | — |
| 2.1 | Tests first | test-engineer | construction/U-NN/test-plan.md + failing tests | — |
| 2.2 | Implement | developer (+ ai-engineer for AI units) | code | — |
| 2.3 | Coding gates CG1–CG14 | code-quality (gate mode) | construction/U-NN/gates.md | auto |
| 2.3f | ↳ on failure: self-heal (≤ 3 attempts/gate) → RCA → escalate to the user | self-heal, rca | self-heal.md, rca/RCA-NN-*.md | auto |
| 2.4 | Verify NFRs (hot-path/critical units) | scalability, robustness (verify mode) | performance.md, robustness.md | CG11/CG12 |
| 2.5 | Code & security review | reviewer | construction/U-NN/review.md | |
| 2.6 | Unit sign-off | orchestrator + user | aidlc-state.md | PG5 |
| OPERATIONS | ||||
| 3.1 | CI/CD (runs all CGs), IaC, observability, runtime self-healing | devops (+ self-heal design) | pipeline, IaC, operations/observability.md | |
| 3.2 | Release readiness | release-manager | operations/release-checklist.md, runbook.md, CHANGELOG.md | PG6 |
| 3.3 | Incidents / escaped bugs (any time) | rca → developer | rca/RCA-NN-*.md + regression test | — |
Agent names are prefixed with aidlc- (e.g. aidlc-code-quality).
Commands the user may say
/aidlcor "start": createaidlc-state.mdfrom the template, ask only for what you can't infer (intent, users, work type, deploy target, PII?, AI?), then begin./aidlc status: readcontext.min.md, then report the phase, the last gate, the next step and the blockers./aidlc gates [U-NN]: run 2.3 for a unit (with the 2.3f loop)./aidlc spike <question>: run SP./aidlc audit: runaidlc-codebase-auditoron the current repo (decisions, drawbacks, edge cases, roadmap)./aidlc rca <incident/bug>: run 3.3.
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.

