source-driven-development
Skill4 starsChanged 3 months agoMITView on GitHub
Ground framework-specific implementation in official documentation — detect stack versions, fetch authoritative docs, implement cited patterns, cite sources. Load when building with a library or framework where API correctness matters, when the user wants verified or documented code, or before writing framework code from memory. Also triggers on "source driven development", "cite the docs", "check official documentation", "verify against docs", "don't hallucinate APIs". Not for pure logic or renames. Pairs with feature-spec and test-driven-development.
72 other projects use a version of this.
Adopted by
Most recently taken up by
- addyosmani/agent-skillstheir own version, first seen Nov 14, 2026
- ishandutta2007/Awesome-Agent-Skillstheir own version, first seen Nov 8, 2026
- TheAstroByte/astrobyte-agent-skillstheir own version, first seen Nov 8, 2026
- MMACABRE/SKILLStheir own version, first seen Nov 8, 2026
- celestialdust/achilles-skillstheir own version, first seen Nov 1, 2026
- steven123397/nexuskit-skillstheir own version, first seen Nov 1, 2026
Used with
Projects with it also use constraint-driven-development, verification-before-completion and 7 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.
MrCipherSmithSkillconstraint-driven-development
Use when a project needs a persistent, measurable quality bar — starting a project without documented standards, 'set up constraints', or when coverage/perf/a11y thresholds get re-debated per-PR. Produces CONSTRAINTS.md with enforceable numbers, ratchets, and a guard against silent weakening. NOT for throwaway prototypes or a one-off CI setup.
Also in 13 projects that use this
OpenAISkillverification-before-completion
Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always
Also in 6 projects that use this
ElasticSkillaccessibility
Accessibility guidance for Kibana. Use this skill when working with or reviewing EUI components, resolving a11y-related (@elastic/eui) ESLint issues, and ensuring proper use of ARIA attributes, focus management, keyboard interactions, and accessible naming conventions.
Also in 3 projects that use this
PostHogSkillkarpathy-guidelines
Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria. Based on https://github.com/multica-ai/andrej-karpathy-skills.
Also in 3 projects that use this
OpenAISkillfinishing-a-development-branch
Use when implementation is complete, all tests pass, and you need to decide how to integrate the work
Also in 3 projects that use this
CloudflareSkillreceiving-code-review
Use when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation
Also in 3 projects that use this
- +2 more from openai
Other versions72 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 72
- addyosmani/agent-skills★ 118ksame name
- shashankswe2020-ux/whoop-mcp★ 181same name
- tranhieutt/software_development_department★ 71same name
- drvoss/everything-copilot-cli★ 49same name
- vinvcn/addyosmani-agent-skills-zh★ 46same name
- jmxt3/gitscape.ai★ 33same name
- sampleXbro/agentsmesh★ 27same name
- jaskirat1616/grok-skills★ 22same name
- ishandutta2007/Awesome-Agent-Skills★ 21same name
- galando/temper★ 18same name
62 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.
In this repository125 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .cursor/rules/agent-loom-routing.mdcCursor rule
- .cursor/rules/agent-loom-skills-index.mdcCursor rule
- adversarial-hatSkill
- agent-builderSkill
- agent-launcherSkill
- agent-loom-syncSkill
- agent-observabilitySkill
- agent-run-retroSkill
- agent-system-architectureSkill
- api-and-interface-designSkill
- api-deprecation-and-migrationSkill
- apply-paper-to-projectSkill
- app-security-hardeningSkill
- architectural-decision-logSkill
- assumption-mappingSkill
- brainstormingSkill
- browser-testing-with-devtoolsSkill
- business-modelingSkill
- ci-cd-and-automationSkill
- codebase-understandingSkill
- code-review-crspSkill
- code-simplificationSkill
- compress-skillSkill
- context-engineeringSkill
- create-agent-promptSkill
- cross-link-skillsSkill
- customer-discoverySkill
- debug-and-fixSkill
- deep-thinkingSkill
- dependency-mappingSkill
- deploy-anywhereSkill
- deprecate-skillSkill
- design-directionSkill
- design-reviewSkill
- design-systemSkill
- dynamic-routingSkill
- eval-judgeSkill
- eval-outputSkill
- eval-pipelineSkill
- eval-rubric-designSkill
- experimentationSkill
- experiment-backlogSkill
- experiment-readoutSkill
- experiment-runbookSkill
- experiment-specSkill
- fault-localizeSkill
- feature-specSkill
- fermiSkill
- first-principlesSkill
- frontend-designSkill
- generate-changelogSkill
- git-workflow-and-versioningSkill
- gsap-animationSkill
- harness-engineeringSkill
- harness-evolutionSkill
- harness-generationSkill
- idea-evaluationSkill
- idea-generationSkill
- implementation-planSkill
- The first 60 shown.
The file
We split this skill into the rules it gives an agent
11 rules. 0 are common, 9 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 65 edited versions we read keep these.
9
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 29
Read dependency manifests (package.json, pyproject.toml, go.mod, etc.) before writing framework-specific code.
- Distinctive: line 31
Never cite Stack Overflow, random blogs, or training data as primary authority.
- Very specific: line 32
If versions are ambiguous, ask once — do not guess.
- Distinctive: line 42
Read the project's dependency file.
- Distinctive: line 50
If versions are missing, ask the user before implementing.
- Distinctive: line 57
If official sources conflict, surface the discrepancy to the user.
- Distinctive: line 61
Use signatures and patterns from the fetched docs.
- Distinctive: line 62
Prefer current recommended APIs over deprecated ones from memory.
- Distinctive: line 63
When docs conflict with existing project code, present options A/B and wait for preference.
- Distinctive: line 76
If no official doc exists for a pattern:
- Distinctive: line 165
No changes — citation audit passed.
“Common” is measured: kept by at least 6 in 10 of the 65 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
Source-Driven Development
You implement framework-specific code from official documentation for the detected version, not from training-data memory. Every non-trivial framework decision is cited or flagged unverified.
Hard Rules
Read dependency manifests (package.json, pyproject.toml, go.mod, etc.) before writing framework-specific code.
Fetch the specific doc page for the feature — not the framework homepage.
Never cite Stack Overflow, random blogs, or training data as primary authority.
If versions are ambiguous, ask once — do not guess.
Surface conflicts between docs and existing codebase; do not silently pick one.
Mark patterns you could not verify as UNVERIFIED explicitly.
Workflow
Step 1 — Detect stack and versions
Read the project's dependency file. State findings explicitly:
STACK DETECTED:
- [package] [version] (from [file])
→ Fetching official docs for [feature].If versions are missing, ask the user before implementing.
Step 2 — Fetch official documentation
Fetch the relevant documentation page for the exact API or pattern. Use references/source-hierarchy.md. Prefer hooks/sdd-cache (Claude Code) or python3 .agents/skills/research-skill/scripts/doc_cache.py "<url>" — see research-skill → references/doc-cache.md.
Extract: API signatures, recommended patterns, deprecations, migration notes. If official sources conflict, surface the discrepancy to the user.
Step 3 — Implement documented patterns
- Use signatures and patterns from the fetched docs.
- Prefer current recommended APIs over deprecated ones from memory.
- When docs conflict with existing project code, present options A/B and wait for preference.
Step 4 — Cite sources
In code (framework-specific lines only):
// useActionState for form pending state — React 19
// Source: https://react.dev/reference/react/useActionState#usageIn chat: full URLs, quote non-obvious passages, note browser/runtime support when relevant.
If no official doc exists for a pattern:
UNVERIFIED: No official documentation found. Based on training data — verify before production.When NOT to use
- Renames, typos, file moves with no API surface change
- Pure language logic identical across versions
- User explicitly wants speed over verification ("just do it quickly")
Gotchas
- Training data lags; "confident" APIs are often deprecated.
- Fetching the whole docs site wastes tokens — one page per decision.
- Simple snippets become copy-paste templates — wrong patterns spread fast.
- Version skew: React 18 patterns in a React 19 repo break silently.
Common Rationalizations
| Excuse | Reality |
| "I'm confident about this API" | Confidence is not evidence. Verify signatures against current docs. |
| "Fetching docs wastes tokens" | One wrong API costs hours of debug time. |
| "Docs won't have what I need" | Absence means the pattern may not be officially recommended — flag it. |
| "I'll add a disclaimer instead" | Either cite or mark UNVERIFIED — hedging helps nobody. |
| "Simple task, no need to check" | Simple wrong patterns get copied across the codebase. |
Output Format
## Source-driven — [feature/slug]
Stack: [detected versions]
Docs fetched: [URLs]
Implementation: [summary]
Citations: [list or UNVERIFIED items]
Conflicts with codebase: [none | options A/B]Examples
<examples>
<example>
<input>Add React 19 form submission with pending state.</input>
<output>
Detect React 19.x from package.json. Fetch react.dev/useActionState.
Implement with cited comment. Reject manual useState+isPending if docs recommend useActionState unless user chooses consistency with existing code.
</output>
</example>
</examples>Verification
- Versions read from dependency file (or user confirmed)
- Official docs fetched for each framework-specific pattern used
- Code matches current-version documented APIs (no deprecated APIs without flag)
- Non-trivial decisions have full URL citations or UNVERIFIED labels
- Doc vs codebase conflicts surfaced to user
Red Flags
- API choice made from training data not fetched docs
- Entire docs site fetched instead of one decision page
- Deprecated API used because snippet was memorable
- Version in code mismatches version cited from docs
Reference Files
references/source-hierarchy.md: Authority order and non-authoritative sources — read at Step 2.
Prune Log
Last pruned: 2026-07-04
- No changes — citation audit passed; content current (improve-skills full pass 2026-07-04)
Impact Report
Feature: [slug] | Stack: [versions]
Docs: [count] fetched | UNVERIFIED: [count]
Conflicts surfaced: [yes/no]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.

