Skip to content
Back to search

dspy-agent-skills

Production-grade DSPy 3.2.x agent skills + validated end-to-end examples for Claude Code and Codex CLI — fundamentals, evaluation, GEPA, BetterTogether, and RLM.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/intertwine--dspy-agent-skills/badge.svg)](/listings/intertwine--dspy-agent-skills)

HTML

<a href="/listings/intertwine--dspy-agent-skills"><img src="/api/listings/intertwine--dspy-agent-skills/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

Claude Code (via marketplace)

/plugin marketplace add intertwine/dspy-agent-skills
/plugin install dspy-agent-skills@dspy-agent-skills

Agent Skills CLI (`npx skills`)

npx skills add intertwine/dspy-agent-skills --list
npx skills add intertwine/dspy-agent-skills --skill '*' -a codex -y

The Vercel `skills` CLI currently expects a GitHub `owner/repo`, URL, well-known HTTPS endpoint, or local path as its source. The bare form `npx skills add dspy-agent-skills` is not resolvable unless the upstream CLI adds a source alias, so use `intertwine/dspy-agent-skills`.

Claude Code + Codex (repo checkout)

git clone https://github.com/intertwine/dspy-agent-skills
cd dspy-agent-skills
./scripts/install.sh           # symlinks into ~/.claude/skills/ and ~/.agents/skills/

Flags: `--claude-only`, `--codex-only`, `--copy` (copy instead of symlink), `--uninstall`, `--dry-run`.

Manual

Drop `skills/*` into `~/.claude/skills/` (Claude Code) or `~/.agents/skills/` (Codex CLI). See [docs/installation.md](docs/installation.md) for all options.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 26 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. {
  2. "name": "dspy-agent-skills",
  3. "owner": {
  4. "name": "Bryan Young",
  5. "email": "bryan@intertwinesys.com"
  6. },
  7. "plugins": [
  8. {
  9. "name": "dspy-agent-skills",
  10. "source": "./",
  11. "description": "Production-grade DSPy 3.2.x skills: fundamentals, evaluation, GEPA, BetterTogether, RLM, and end-to-end workflow.",
  12. "version": "0.2.3",
  13. "author": {
  14. "name": "Bryan Young",
  15. "email": "bryan@intertwinesys.com"
  16. },
  17. "homepage": "https://github.com/intertwine/dspy-agent-skills",
  18. "repository": "https://github.com/intertwine/dspy-agent-skills",
  19. "license": "MIT",
  20. "keywords": ["dspy", "gepa", "rlm", "optimization", "llm", "evaluation"],
  21. "category": "llm-frameworks",
  22. "tags": ["dspy", "optimization", "evaluation"],
  23. "strict": true
  24. }
  25. ]
  26. }
README.md · README · 146 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # DSPy Agent Skills
  2. [![DSPy 3.2.x](https://img.shields.io/badge/DSPy-3.2.x-0A7B83)](https://dspy.ai/)
  3. **Production-grade DSPy 3.2.x skills for coding agents.** A synthesized, spec-compliant pack of five agent skills that turns Claude Code, Codex CLI, and any other [agentskills.io](https://agentskills.io)-compatible agent into a DSPy expert.
  4. - ✅ Validated against DSPy 3.2.1 (the real API, not inferred from stale docs)
  5. - ✅ Single source of truth for both **Claude Code** and **Codex CLI**
  6. - ✅ Progressive disclosure (short `SKILL.md` + deep `reference.md`)
  7. - ✅ Runnable `example_*.py` scripts with offline `--dry-run`
  8. - ✅ Includes a DSPy 3.2.x `BetterTogether` chaining example
  9. - ✅ Plugin manifest + marketplace manifest for one-click install
  10. - ✅ Validation tests for frontmatter spec, JSON schema, Python AST, skill-doc correctness, and version alignment
  11. ## What's inside
  12. | Skill | When it auto-invokes |
  13. |---|---|
  14. | [`dspy-fundamentals`](skills/dspy-fundamentals/SKILL.md) | Any new DSPy code: Signatures, Modules, Predict/ChainOfThought/ReAct, save/load |
  15. | [`dspy-evaluation-harness`](skills/dspy-evaluation-harness/SKILL.md) | Writing metrics, splitting dev/val sets, calling `dspy.Evaluate` |
  16. | [`dspy-gepa-optimizer`](skills/dspy-gepa-optimizer/SKILL.md) | Optimizing/compiling DSPy programs with `dspy.GEPA` |
  17. | [`dspy-rlm-module`](skills/dspy-rlm-module/SKILL.md) | Long context, codebase QA, recursive exploration via `dspy.RLM` |
  18. | [`dspy-advanced-workflow`](skills/dspy-advanced-workflow/SKILL.md) | End-to-end builds — orchestrates the other four |
  19. ## Install
  20. ### Claude Code (via marketplace)
  21. ```text
  22. /plugin marketplace add intertwine/dspy-agent-skills
  23. /plugin install dspy-agent-skills@dspy-agent-skills
  24. ```
  25. ### Agent Skills CLI (`npx skills`)
  26. ```bash
  27. npx skills add intertwine/dspy-agent-skills --list
  28. npx skills add intertwine/dspy-agent-skills --skill '*' -a codex -y
  29. ```
  30. The Vercel `skills` CLI currently expects a GitHub `owner/repo`, URL, well-known HTTPS endpoint, or local path as its source. The bare form `npx skills add dspy-agent-skills` is not resolvable unless the upstream CLI adds a source alias, so use `intertwine/dspy-agent-skills`.
  31. ### Claude Code + Codex (repo checkout)
  32. ```bash
  33. git clone https://github.com/intertwine/dspy-agent-skills
  34. cd dspy-agent-skills
  35. ./scripts/install.sh # symlinks into ~/.claude/skills/ and ~/.agents/skills/
  36. ```
  37. Flags: `--claude-only`, `--codex-only`, `--copy` (copy instead of symlink), `--uninstall`, `--dry-run`.
  38. ### Manual
  39. Drop `skills/*` into `~/.claude/skills/` (Claude Code) or `~/.agents/skills/` (Codex CLI). See [docs/installation.md](docs/installation.md) for all options.
  40. ## Five-second demo
  41. In your agent, say:
  42. > "Build a DSPy sentiment classifier, optimize it with GEPA, and save the artifact."
  43. The agent auto-loads `dspy-advanced-workflow`, which chains the other skills and outputs a full baseline → GEPA → export pipeline. No further prompting needed.
  44. ## End-to-end examples (current committed artifacts)
  45. Three runnable demos under [`examples/`](examples/) exercise every skill against real LMs and ship with **committed baseline vs. GEPA-optimized numbers** plus explicit `3.1.3` vs. `3.2.0` comparison notes.
  46. | Example | Artifact DSPy | Task LM | Baseline | Optimized | Δ | Status |
  47. |---|---|---|---:|---:|---:|---|
  48. | [01-rag-qa](examples/01-rag-qa/) | 3.2.0 | Ministral 3B 2512 | 80.47 | **100.00** | **+19.53** | Clean comparison refreshed on 2026-04-28 |
  49. | [02-math-reasoning](examples/02-math-reasoning/) | 3.2.0 | Ministral 3B 2512 | 85.00 | **93.33** | **+8.33** | Refreshed on 2026-04-21 |
  50. | [03-invoice-extraction](examples/03-invoice-extraction/) | 3.1.3 | Liquid LFM 2.5 1.2B (free) | 0.833 | **0.931** | **+0.098** | Historical artifact retained |
  51. The refreshed `01` and `02` artifacts use the paid pair `openrouter/mistralai/ministral-3b-2512` + `openrouter/qwen/qwen3-30b-a3b-instruct-2507`. `03` stays on its historical DSPy `3.1.3` artifact because a clean DSPy `3.2.0` baseline on the same Liquid/Nemotron pair already reached `0.944`, leaving little useful headroom for a replacement GEPA artifact. See [`examples/README.md`](examples/README.md) and each example's `version_comparison.md` for the exact commands and caveats.
  52. ## Grounding
  53. Every API claim is grounded in:
  54. - https://dspy.ai/ (official docs, DSPy 3.2.x)
  55. - https://code.claude.com/docs/en/skills.md (Claude Code skill spec)
  56. - https://developers.openai.com/codex/skills (Codex skill spec)
  57. ## Development
  58. ```bash
  59. # Run validation suite
  60. uv run --with pytest python -m pytest tests/ -v
  61. # Smoke-test every example offline (no API key needed)
  62. for f in skills/*/example_*.py; do uv run --with dspy python "$f" --dry-run; done
  63. # Validate the current DSPy API surface used by these skills
  64. env -u UV_EXCLUDE_NEWER uv run --with dspy==3.2.1 python scripts/check_dspy_surface.py
  65. # Live GEPA run (requires OPENAI_API_KEY)
  66. cd skills/dspy-advanced-workflow
  67. OPENAI_API_KEY=... uv run --with dspy python example_pipeline.py --auto light
  68. ```
  69. If `uv run --with dspy` resolves an older DSPy release instead of the current `3.2.1` wheel, check whether `UV_EXCLUDE_NEWER` or a stale package mirror is hiding the new release. The exact 3.2.1 override we validated for this repo is:
  70. ```bash
  71. env -u UV_EXCLUDE_NEWER uv run --with dspy==3.2.1 python -c 'import dspy; print(dspy.__version__)'
  72. ```
  73. ## Compatibility
  74. - **DSPy**: 3.2.x (tested against 3.2.1; committed example artifacts remain explicitly labeled by the DSPy version that produced them)
  75. - **Claude Code**: current (skill spec as of 2026-04-17)
  76. - **Codex CLI**: current Agent Skills format
  77. - **Python**: 3.10+
  78. - **Deno**: required only for `dspy.RLM` examples (Pyodide sandbox)
  79. ## Layout
  80. ```
  81. dspy-agent-skills/
  82. ├── .claude-plugin/
  83. │ ├── plugin.json
  84. │ └── marketplace.json
  85. ├── skills/
  86. │ ├── dspy-fundamentals/{SKILL.md, reference.md, example_qa.py}
  87. │ ├── dspy-evaluation-harness/{SKILL.md, reference.md, example_metric.py}
  88. │ ├── dspy-gepa-optimizer/{SKILL.md, reference.md, example_gepa.py}
  89. │ ├── dspy-rlm-module/{SKILL.md, reference.md, example_rlm.py}
  90. │ └── dspy-advanced-workflow/{SKILL.md, reference.md, example_pipeline.py}
  91. ├── scripts/install.sh # dual-target installer
  92. ├── tests/ # spec validators
  93. ├── docs/{installation,usage,CHANGELOG}.md
  94. ├── README.md LICENSE .gitignore
  95. ```
  96. ## Version
  97. **v0.2.3** • Targets DSPy 3.2.x
  98. ## License
  99. MIT — see [LICENSE](LICENSE).
  100. ## Credits
  101. Draft contributors: Bryan Young ([@intertwine](https://github.com/intertwine)) with Grok (xAI).
  102. Validation, spec-alignment, and dual-agent packaging: Claude Opus 4.7, April 2026.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.