Skip to content
SkillModel-led2,066 tokens in use

source-driven-development

Ground every framework decision in official documentation, cite the source and flag what could not be verified.

119,377 stars on the repository MIT Licenseaddyosmani/agent-skills · skills/source-driven-development
npx skills add addyosmani/agent-skills --skill source-driven-development
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing source-driven-development costs

56tokens idle
2,066tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

Nothing detected: no shell, network, file writes, git or credentials in its files.

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Excellent
  • Frontmatter12/15
  • Structure17/25
  • Examples15/15
  • Output15/15
  • Size15/15
  • Resources10/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Excellent
  • Development activity96
  • Maintenance100
  • Community76
  • Maturity32
  • Momentum40

Source-Driven Development has an agent verify framework-specific code against official documentation. It follows four steps: detect the stack and exact versions from the dependency file, fetch the specific documentation page, implement the documented pattern and cite the source with a full URL.

The skill ranks sources by authority, excludes Stack Overflow, blog posts and the model’s own memory as primary sources, surfaces conflicts between docs and existing code, and treats fetched pages as data that never directs the agent.

What You Get

  • A four-step process: detect, fetch, implement and cite
  • A ranked list of authoritative and non-authoritative sources
  • Citation formats for code comments and conversation
  • A marker for unverified patterns, red flags and a verification checklist

Common Use Cases

  • Building with a framework where the correct pattern depends on the version
  • Creating boilerplate that others will copy
  • Checking code for deprecated APIs
  • Producing code the user can verify against sources

What Makes It Unique

The skill is explicit that the agent’s own training data is not an acceptable primary source and requires an UNVERIFIED label when no documentation is found. It also includes retrieval safety rules, which keep fetched documentation from steering the agent or adding outbound endpoints to generated code.

Under The Hood

The skill is a single SKILL.md of instructions. The agent reads the project’s dependency files and then uses a web fetch tool to retrieve official documentation pages over the network. It runs no scripts of its own and points to the security-and-hardening skill for the prompt injection threat model.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers