project-development
Skill0 starsChanged 4 months agoMITView on GitHub
Master specification for reviewing, planning, and executing software projects. Provides the canonical structure, mandatory practices, and role-specific rules. Use when performing any action on a project.
11 other projects use a version of this.
Adopted by
Most recently taken up by
- docxology/templatetheir own version, first seen Oct 30, 2026
- AbdelrhmanAlquran/claude-skillstheir own version, first seen Oct 20, 2026
- xfstudio/skillstheir own version, first seen Oct 13, 2026
- brake71/1c-ssl-skillstheir own version, first seen Oct 13, 2026
- neilyonglu/project-developmenttheir own version, first seen Oct 11, 2026
- muratcankoylan/Agent-Skills-for-Context-Engineeringtheir own version, first seen Aug 10, 2026
Used with
Projects with it also use self-improvement-loops, long-horizon-prompting and 3 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
DVwebBiz11Skillself-improvement-loops
This skill should be used when the harness, scaffold, workflow, or optimizer itself is the optimization target: recursive self-improvement (RSI) loops, meta-harnesses, self-improving harnesses that mine their own failures and propose bounded edits, evolutionary or population-based search over agent scaffolds, acceptance gates for self-modifying systems, and agentic context evolution where the mechanism that produces context is versioned and evolved. Route governance of a single autonomous loop (locked surfaces, durable logs, rollback, novelty gates, approval boundaries) to harness-engineering, measurement and quality-gate design to evaluation, judge design to advanced-evaluation, and remote sandbox infrastructure to hosted-agents.
Also in 4 projects that use this
DVwebBiz11Skilllong-horizon-prompting
This skill should be used when writing, enhancing, or evaluating the launch prompt for a long-running autonomous agent or a parallel multi-agent orchestration attacking a hard problem: pseudo-formal task briefs that define terms and an exact success predicate linguistically, enumerate non-counting outcomes, set persistence rules with explicit stop and return conditions and effort floors, manage a diverse portfolio of parallel approaches with an approach registry and blocked-route bookkeeping, and gate the return on adversarial audit. Route agent topology and coordination protocols to multi-agent-patterns, runtime control surfaces and loop governance to harness-engineering, evaluator and quality-gate construction to evaluation, judge design to advanced-evaluation, and compaction or memory mechanics to context-compression and memory-systems.
Also in 4 projects that use this
AbdelrhmanAlquranSkilllatent-briefing
This skill should be used when the user asks to "share memory between agents", "KV cache compaction for multi-agent", "orchestrator worker context", "latent briefing", "reduce worker tokens", "cross-agent memory without summarization", or discusses Attention Matching compaction, recursive language models with workers, or token explosion in hierarchical agents.
Also in 4 projects that use this
GitHubSkillharness-engineering
Adopt repository-level harness engineering for coding agents. Use when a user wants to prevent repeated AI coding-agent mistakes by turning failures into durable instructions, drift checks, regression tests, failure memory, and adoption reports tailored to the target repository.
Also in 5 projects that use this
BlockSkillfrontend-design
Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, or applications. Generates creative, polished code that avoids generic AI aesthetics.
Also in 3 projects that use this
Other versions11 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 11
- muratcankoylan/Agent-Skills-for-Context-Engineering★ 18ksame name
- brake71/1c-ssl-skills★ 85same name
- docxology/template★ 23same name
- xfstudio/skills★ 9same name
- fqmyysjjd/agent-feed★ 4same name
- gsmlg-app/flutter-app-template★ 2same name
- Downey6689/skills★ 1same name
- goodnessibeh/ai-dev-boilerplate★ 1same name
- AbdelrhmanAlquran/claude-skills★ 0same name
- DVwebBiz11/claude-skills★ 0same name
1 more not listed
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
Show its rules and text
We split this skill into the rules it gives an agent
21 rules. 0 are common, 5 are distinctive, and 16 are very specific to this project.
0
Common
Most of the 10 edited versions we read keep these.
5
Distinctive
This author's own choices. Worth a look.
16
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 13
Documentation is Paramount: Every significant decision, requirement, and procedure must be written down in the prescribed location.
- Very specific: line 89
Refinement Rules: When a FEAT is started, refine it into one or more functional requirements (FR) and append them to 01-REQUIREMENT.md with …
- Very specific: line 91
Key Rule: Do not split or refine FEAT/IDEA items in this file.
- Very specific: line 114
Create the ADR design file in docs/02-DESIGN/ when the decision is made.
- Very specific: line 115
When implementation completes, write an implementation summary and archive it to docs/08-ARCHIVE/ (e.g., ADR-007-IMPLEMENTATION-SUMMARY.md).
- Distinctive: line 116
Add an "实施记录与归档" (Implementation Records & Archive) section at the end of the ADR design file, linking to the archived summary.
- Very specific: line 117
Update docs/02-DESIGN/INDEX.md with the ADR entry and an archive index table.
- Very specific: line 176
Environment Variables: All required ENV vars, with examples or links to vault.
- Very specific: line 212
When items in PROJECT_BOARD.md are completed, move them to 08-ARCHIVE/HISTORY.md and remove from the board.
- Distinctive: line 217
Preferred: Update it whenever you add a new task, or finish a task.
- Very specific: line 224
Completed Items: When a task is done, move its entry to docs/08-ARCHIVE/HISTORY.md (do not leave a stale "Completed" section in the board).
- Very specific: line 226
FEAT/IDEA Items: Do not maintain FEAT-xxx / IDEA-xxx tables in PROJECT_BOARD.md — they live in docs/00-ROADMAP.md.
- Distinctive: line 240
Pure unit tests: test a single class/function with all dependencies (DB, network, filesystem) mocked or stubbed.
- Distinctive: line 242
Acceptance tests: verify real user journeys end-to-end, including real external systems (HTTP APIs, real DB).
- Very specific: line 244
pytest.ini: must set addopts = --import-mode=importlib to allow same-basename test files across unit/, functional/, and acceptance/ … - Very specific: line 245
Test file naming: should reflect the functionality/scenario being tested (e.g., test_holdings_backtest_data_validation.py), not the …
- Very specific: line 271
Check the current state in PROJECT_BOARD.md.
- Very specific: line 281
Always update PROJECT_BOARD.md when work items are created or status are changed.
- Very specific: line 285
Place tests in the correct layer: tests/unit/, tests/functional/, or tests/acceptance/ (see §Test Layering above).
- Very specific: line 287
Ensure pytest.ini has --import-mode=importlib enabled.
- Very specific: line 288
After test changes, update docs/07-TEST_RESULTS.md with new counts and per-fix coverage mapping.
“Common” is measured: kept by at least 6 in 10 of the 10 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Software Project Development Master Skill
This skill defines the single source of truth for how this AI agent team understands, develops, and delivers software projects. It mandates a consistent structure and a documentation-first workflow to ensure clarity, reproducibility, and effective collaboration.
Core Principles (All Roles)
- Documentation is Paramount: Every significant decision, requirement, and procedure must be written down in the prescribed location.
- Traceability: Code and features must be traceable to a source requirement (
docs/01-REQUIREMENT.md) or a recorded decision (docs/02-DESIGN/). - Reproducibility: Any team member must be able to build, test, and deploy the project using only the instructions in this repository.
- Explicit State: The current state of work, bugs, and ideas must be visibly tracked, ideally in a synchronized project management system.
Project Structure & Canonical Documents
Each project shall have a top directory under teamshared/ directory, e.g. teamshared/<project_name>/. Under the top directory, follow below rules to manage project files.
The following directory and file structure is mandatory. All paths are relative to the project root (./).
./ ├── README.md # Project's front page and main index. ├── docs/ # Primary documentation directory. │ ├── 00-ROADMAP.md # Product roadmap: FEAT-xxx (enhancements) & IDEA-xxx (future). │ ├── 01-REQUIREMENT.md # All user requirements: functional (FR), non-functional (NFR), + FEAT/IDEA refs. │ ├── 02-DESIGN/ # Architecture and decision records. │ │ ├── INDEX.md # Index of all Architectural Decision Records (ADRs) + archive index. │ │ └── ADR-001-title.md # Individual ADR (see format below). │ ├── 03-TEST.md # Acceptance criteria, test plan, and results. │ ├── 04-DEPLOY.md # Step-by-step build, deployment, and ops procedures. │ ├── 05-USER_DOCUMENT.md # End-user guide, API reference, or manual. │ ├── 06-DOMAIN_SPECIFIC/ # (Optional) Domain-specific module documentation. │ ├── 07-TEST_RESULTS.md # Latest test results report (counts, coverage, per-fix mapping). │ ├── 08-ARCHIVE/ # Completed milestones, ADR implementation summaries, historical reports. │ │ └── HISTORY.md # Archived completed items (moved from PROJECT_BOARD.md). │ └── references/ # (Optional) Extended reference docs (e.g., architecture diagrams, legacy design notes). ├── src/ # Application source code. ├── tests/ # Automated test suites (layered - see §Test Layering below). │ ├── unit/ # Pure unit tests: single class/function, all deps mocked. │ ├── functional/ # Multi-component wiring: CLI end-to-end, real in-memory DB, mocked boundary. │ ├── acceptance/ # End-to-end with real network/DB/external systems. │ └── pytest.ini # Test config (MUST set --import-mode=importlib for layered tests/). ├── scripts/ # Build, utility, and environment scripts. ├── PROJECT_BOARD.md # Active work tracking (completed items archived to 08-ARCHIVE/HISTORY.md). ├── MONITORING_STATUS.md # (Optional) Data source / system health monitoring status. ├── SUGGESTIONS.md # (Optional) User/team suggestions and feedback log. └── [dependency files] # e.g., requirements.txt, package.json, pyproject.toml.
(*): File may live in a different standard location for the project's language/framework.
Numbering Convention for docs/
Documents under docs/ use a zero-padded NN- prefix to enforce a logical reading order:
| Prefix | Purpose | Mutability |
00- | Roadmap (entry point) | Updated when FEAT/IDEA status changes |
01- | Requirements (the "Why") | Updated when FRs are added/refined |
02- | Design (the "How" + "Why this way") | Updated when ADRs are added/changed |
03- | Test strategy | Updated when test approach changes |
04- | Deployment runbook | Updated when deploy process changes |
05- | User documentation | Updated when UX changes |
06- | (Optional) Domain-specific modules | As needed |
07- | Test results report | Updated per test run / milestone |
08- | Archive (completed/historical) | Append-only; never edit archived content |
1. README.md - The Project Index
Purpose: The first point of entry. It must orient a new developer or user in under 5 minutes. Content Must:
- Project Name & Brief Description: What it is, in one paragraph.
- Current Status: "Active / In Planning / Archived". Link to the main
PROJECT_BOARD.md. - Quick Start: A 3-5 command sequence to get a development environment running.
- Full Document Index: Direct links to every document in the
docs/directory. - Team & Contact: Point of contact for questions.
2. docs/00-ROADMAP.md - The "Where We're Going"
Purpose: Track product roadmap items that are not yet refined into functional requirements. Content Must:
- Completed Milestones Overview: High-level summary table (details go to
08-ARCHIVE/HISTORY.md). - FEAT-xxx Items: Planned feature enhancements (e.g.,
FEAT-001Parameter Optimization). Include status and estimated effort. - IDEA-xxx Items: Long-term, unprioritized ideas (e.g.,
IDEA-001Multi-data-source support). - Refinement Rules: When a FEAT is started, refine it into one or more functional requirements (FR) and append them to
01-REQUIREMENT.mdwith a参考: FEAT-xxxannotation.
Key Rule: Do not split or refine FEAT/IDEA items in this file. Only refine them when the item is actually being started, and the refined FRs go to 01-REQUIREMENT.md, not back here.
3. docs/01-REQUIREMENT.md - The "Why"
Purpose: Define the problem, stakeholders, goals, and what "done" looks like. This file contains all user requirements. Content Must:
- Problem Statement: The pain point or opportunity this project addresses.
- Goals & Non-Goals: Clear, measurable objectives and explicit exclusions.
- User Personas & Stories: Who uses this and what they need to accomplish.
- Functional Requirements (FR-xxx): Specific capabilities, each with a unique ID. When an FR is refined from a FEAT, annotate with
参考: FEAT-xxx. - Non-Functional Requirements (NFR): Performance, security, reliability needs.
- FEAT/IDEA Reference Index: A section (e.g., §10) listing all FEAT-xxx and IDEA-xxx IDs with links to
00-ROADMAP.mdfor details.
Key Rule: FEAT-xxx and IDEA-xxx items themselves do not live here — only their refined FRs (with back-references) and a reference index.
4. docs/02-DESIGN/ - The "How" and "Why This Way"
Purpose: Record the significant architectural and high-level design decisions. Process: Major decisions are recorded as Architectural Decision Records (ADRs).
- File Naming:
ADR-001-use-python-fastapi.md - ADR Lifecycle:
- Create the ADR design file in
docs/02-DESIGN/when the decision is made. - When implementation completes, write an implementation summary and archive it to
docs/08-ARCHIVE/(e.g.,ADR-007-IMPLEMENTATION-SUMMARY.md). - Add an "实施记录与归档" (Implementation Records & Archive) section at the end of the ADR design file, linking to the archived summary.
- Update
docs/02-DESIGN/INDEX.mdwith the ADR entry and an archive index table.
- Create the ADR design file in
ADR Format (Template):
# [ADR-001] Use Python and FastAPI Status: [Proposed | Accepted | Superseded by ADR-002] Deciders: [Architect Agent, Developer Agent] Date: 2023-10-27 ## Context and Problem Statement [The motivating context and the problem we are solving.] ## Decision Drivers [Driver 1, e.g., Development speed] [Driver 2, e.g., Team familiarity] ## Considered Options [Option 1] [Option 2] [Option 3, including the chosen one] ## Decision Outcome Chosen option: "[Option 3]", because... ## Consequences Positive: [e.g., Rapid prototyping, strong async support.] Negative: [e.g., Python's runtime performance trade-off.] ## Implementation Records & Archive > Implementation summaries are archived in [../08-ARCHIVE/](../08-ARCHIVE/). | Document | Description | Status | |----------|-------------|--------| | [ADR-001-IMPLEMENTATION-SUMMARY.md](../08-ARCHIVE/ADR-001-IMPLEMENTATION-SUMMARY.md) | Implementation completion summary | ✅ Completed (YYYY-MM-DD) |
5. docs/03-TEST.md - The "Verification"
Purpose: Define how we verify the system works as required. Content Must:
- Testing Strategy: Unit, functional, acceptance testing approach (see §Test Layering below).
- Acceptance Test Matrix: Links requirements (FR-xxx) to specific test cases/scenarios.
- Results & Status: Current test pass/fail status. Detailed results go in
07-TEST_RESULTS.md.
6. docs/04-DEPLOY.md - The "Delivery"
Purpose: Provide complete, copy-pasteable instructions to go from source to a running system. Content Must: Be a runbook. Assume the person reading it has the repo but no other context.
- Prerequisites: OS, installed tools (Docker, Node, Python) with version numbers.
- Environment Variables: All required
ENVvars, with examples or links to vault. - Step-by-Step Commands: For local dev setup, staging, and production deployment.
- Health Checks: How to verify the deployment was successful.
7. docs/05-USER_DOCUMENT.md - The "User Manual"
Purpose: Teach an end-user how to achieve their goals with the system. Content Must be tailored:
- For Applications: Tutorials, how-to guides, screenshot walkthroughs.
- For Libraries/APIs: Getting started, comprehensive API reference, examples.
8. docs/07-TEST_RESULTS.md - The "Test Report"
Purpose: Provide a living report of the latest test results, separate from the test strategy in 03-TEST.md.
Content Must:
- Overview: Total test count, pass/fail counts, test file count.
- Test File Listing: Organized by layer (unit/functional/acceptance) with per-file test counts.
- Bug/Fix Tracking: A table mapping each fixed bug (e.g., Section A/B items) to the specific test class(es) and test count that verify the fix.
- Coverage Gaps: Known untested areas.
- Run Commands: How to run each test layer independently.
9. docs/08-ARCHIVE/ - The "Memory"
Purpose: Consolidate all completed and historical documentation to keep living docs (docs/ root) clean.
Content Includes:
HISTORY.md— Completed milestones and items moved fromPROJECT_BOARD.md.ADR-xxx-IMPLEMENTATION-SUMMARY.md— ADR implementation summaries (linked from the ADR design file in02-DESIGN/).ADR-xxx-P0-FIXES-SUMMARY.md— Bug fix reports tied to an ADR.- Historical review/report files (e.g.,
REVIEW.md,phase1_completion_report.md,P0_FIXES_SUMMARY.md).
Key Rules:
- This directory is append-only: never edit or delete archived content (use new files for corrections if needed).
- Archived ADR summaries must be linked from their parent ADR design file in
02-DESIGN/via an "实施记录与归档" section. - When items in
PROJECT_BOARD.mdare completed, move them to08-ARCHIVE/HISTORY.mdand remove from the board.
10. PROJECT_BOARD.md - The "Living State"
Primary Rule: Do not let this file become stale. It is the heartbeat of the project. It must contain only active/incomplete tasks — completed items are archived to docs/08-ARCHIVE/HISTORY.md.
Preferred: Update it whenever you add a new task, or finish a task. It must be strictly maintained in the following sections:
## In Progress(Current sprint/cycle work)## Backlog(Prioritized future work)## Blocked(Work waiting on external input)## Ideas & Future(Unprioritized ideas — or reference00-ROADMAP.mdfor FEAT/IDEA items)
Completed Items: When a task is done, move its entry to docs/08-ARCHIVE/HISTORY.md (do not leave a stale "Completed" section in the board). Keep only a brief pointer if needed.
FEAT/IDEA Items: Do not maintain FEAT-xxx / IDEA-xxx tables in PROJECT_BOARD.md — they live in docs/00-ROADMAP.md. The board may contain a one-line pointer to the roadmap.
Test Layering (tests/)
Tests are organized into three layers by scope and dependency isolation. Each layer has a distinct purpose and run profile.
| Layer | Directory | Scope | External Deps | Speed | Run When |
| Unit | tests/unit/ | Single class/function | All mocked | Fast (<1s/test) | Every commit |
| Functional | tests/functional/ | Multi-component wiring (CLI→Service→DAO) | Boundary mocked (DB/network), real in-memory DB | Medium | Every commit |
| Acceptance | tests/acceptance/ | End-to-end user scenarios | Real network, DB, file system | Slow, potentially flaky | Pre-release / manual |
Rules
- Pure unit tests test a single class/function with all dependencies (DB, network, filesystem) mocked or stubbed. They must be deterministic and never touch the network.
- Functional tests exercise multiple real components together (e.g., CLI via
CliRunner→ Service → DAO with in-memory SQLite). Mock only at the system boundary (external APIs, real network). - Acceptance tests verify real user journeys end-to-end, including real external systems (HTTP APIs, real DB). They may be flaky and should be run manually or nightly, not on every commit.
- Mixed test files (containing both unit and functional tests) must be split into separate files in the appropriate subdirectory. Use the same basename across directories if needed.
pytest.inimust setaddopts = --import-mode=importlibto allow same-basename test files acrossunit/,functional/, andacceptance/subdirectories without__init__.pyconflicts.- Test file naming should reflect the functionality/scenario being tested (e.g.,
test_holdings_backtest_data_validation.py), not the development phase.
Running Tests by Layer
# Unit only (fast, deterministic, no network)
pytest tests/unit/
# Functional (real DB, CLI end-to-end, no network)
pytest tests/functional/
# Acceptance (real network — run manually/pre-release)
pytest tests/acceptance/
# All
pytest tests/How to Use This Skill: Agent Workflows
When starting work or reviewing a project:
- Always begin by reading the
README.md. - Understand the "Where we're going" by reviewing
docs/00-ROADMAP.md(FEAT/IDEA items, milestones). - Understand the "Why" by reviewing
docs/01-REQUIREMENT.md. - Understand the "How" by reviewing the latest ADRs in
docs/02-DESIGN/. - Check the current state in
PROJECT_BOARD.md. - For historical context, review
docs/08-ARCHIVE/HISTORY.md.
When implementing a new feature or fix:
Depend on your role, check your corresponding rules:
- PJM: pjm.md
- Architect: architect.md
- Developer: developer.md
Always update PROJECT_BOARD.md when work items are created or status are changed. When a task is completed, move it to docs/08-ARCHIVE/HISTORY.md.
When writing or organizing tests:
- Place tests in the correct layer:
tests/unit/,tests/functional/, ortests/acceptance/(see §Test Layering above). - Split mixed test files by type — never mix unit and functional tests in the same file.
- Ensure
pytest.inihas--import-mode=importlibenabled. - After test changes, update
docs/07-TEST_RESULTS.mdwith new counts and per-fix coverage mapping.
When stuck or planning:
- Review relevant ADRs in
docs/02-DESIGN/for context (check the "实施记录与归档" section for archived summaries). - Log the blocker in
PROJECT_BOARD.mdunder "Blocked", or log the idea indocs/00-ROADMAP.mdunder FEAT/IDEA items. - For completed work context, consult
docs/08-ARCHIVE/HISTORY.md.
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.

