Request or Build a Missing Skill

The user needs a DSPy capability that doesn't have a skill yet. Help them contribute it or request it.

This skill always ends with a concrete action on GitHub:

Path A (Build): Create the skill files, commit, push, and open a pull request to lebsral/DSPy-Programming-not-prompting-LMs-skills
Path B (Request): File a GitHub issue on lebsral/DSPy-Programming-not-prompting-LMs-skills describing what's needed

Do not stop at "here's what the PR/issue would look like" — actually create it using gh pr create or gh issue create.

Step 1: Confirm the gap

If $ARGUMENTS is provided, use it. Otherwise ask: "What DSPy capability do you need that isn't covered by an existing skill?"

Verify this is something DSPy actually supports. If it's outside DSPy's scope entirely (e.g., "build a React frontend"), say so and suggest appropriate tools instead.

Check the existing skills in skills/ or ai-* to make sure there isn't already a skill that covers this. If there's a close match, suggest it instead.

Summarize back to the user:

What they need: one sentence
DSPy features involved: which DSPy modules, integrations, or patterns are relevant
Closest existing skill: what's close but doesn't quite fit

Step 2: Choose a path

Ask the user:

Would you like to:

Build the skill — I'll help you create it with proper testing and prepare a PR

Request the skill — I'll draft a GitHub issue so the maintainers know it's needed

Path A: Build the skill

Use skill-creator if available

Check whether the /skill-creator skill is available (it's from the anthropics/skills repo). If available, delegate to it — it handles the full create-test-iterate workflow including evaluation, benchmarking, and description optimization.

When delegating to /skill-creator, first read the repo standards so the skill matches conventions:

Read docs/skill-standards.md — the full authoring checklist (naming, descriptions, gotchas, cross-refs, progressive disclosure, provider-agnostic code)
Read docs/skills-spec.md — the Claude Code skills format (frontmatter fields, file structure, supporting files)
Read CLAUDE.md — repo-level conventions (dual naming, web-developer language, 500-line limit)

Pass the contents of these files as context to /skill-creator so it generates a skill that matches repo standards on the first pass.

Then let skill-creator run its workflow: draft → test → review → iterate → package.

If skill-creator is NOT available

Build the skill manually. First read the standards:

Read docs/skill-standards.md for the full authoring checklist
Read docs/skills-spec.md for the Claude Code skills format
Read 2-3 existing skills in skills/ to match tone and structure

Install skill-creator for future use:

npx skills add anthropics/skills/skill-creator

Write the SKILL.md

Follow the structure in docs/skill-standards.md. Key points:

Description: <WHAT>. Use when <triggers>. Also: <more triggers>. — plain YAML scalar, no quotes
Body: Step 1 gathers context (2-4 questions) → Steps 2-4 core work → Gotchas (3-5) → Cross-references → Additional resources
Code: provider-agnostic with # or ... comment, copy-pasteable with imports
Standalone: every skill must be self-contained — include a reference.md for dspy- skills or ai- skills that heavily reference DSPy APIs
Verify API usage against https://dspy.ai/api/ for correct patterns

Test the skill

Create 2-3 test prompts — realistic requests a developer would make. Run them with the skill active and verify the outputs make sense.

Submit the PR

After creating (and optionally testing) the skill files, submit a pull request. Do all of these steps — don't stop at "here's what to do":

Update README.md — add a row to the problem catalog table in the appropriate position
Create a branch: git checkout -b add-ai-<problem-name>
Stage and commit: git add skills/ai-<problem-name>/ README.md && git commit -m "Add ai-<problem-name> skill"
Push: git push -u origin add-ai-<problem-name>
Open the PR:

gh pr create \
  --repo lebsral/DSPy-Programming-not-prompting-LMs-skills \
  --title "Add ai-<problem-name> skill" \
  --body "$(cat <<'EOF'
## Summary
- **Problem**: <what the user is solving>
- **DSPy features**: <modules, patterns used>
- **Example invocation**: `/ai-<problem-name> <example prompt>`

## Files
- `skills/ai-<problem-name>/SKILL.md` — main instructions
- `skills/ai-<problem-name>/examples.md` — worked examples
- `README.md` — catalog table updated
EOF
)"

Return the PR URL to the user when done.

Path B: Request the skill

File a GitHub issue on the repo. Do not just draft it — actually submit it:

gh issue create \
  --repo lebsral/DSPy-Programming-not-prompting-LMs-skills \
  --title "Skill request: ai-<problem-name>" \
  --assignee lebsral \
  --body "$(cat <<'EOF'
## Problem
<What the user is trying to do, in their words>

## DSPy capability
<Which DSPy modules, integrations, or patterns would power this>

## Example use case
<A concrete scenario where this skill would help>

## Suggested trigger phrases
<2-3 phrases a developer might say that should route to this skill>
EOF
)"

Return the issue URL to the user when done.

Quality checklist

Validate against docs/skill-standards.md (the authoritative checklist). Key items:

[ ] Description: plain YAML scalar, trigger phrases users would actually say
[ ] SKILL.md under 500 lines with progressive disclosure structure
[ ] Gotchas section with 3-5 Claude/DSPy-specific failure modes
[ ] Cross-references with install hint blockquote and /ai-do back-link
[ ] Code examples provider-agnostic with # or ... comment
[ ] reference.md for dspy- skills (or ai- skills that heavily use DSPy APIs)
[ ] README.md catalog table updated with new row
[ ] /ai-do catalog updated with new routing entry

Gotchas

Writing descriptions wrapped in quotes. YAML descriptions must be plain scalars — no double quotes, no apostrophes. Claude defaults to quoting strings but this causes parse failures or ambiguity. Write description: Monitor AI quality... not description: "Monitor AI quality...".
Forgetting to update /ai-do routing table. New skills are invisible if /ai-do cannot route to them. Every PR adding a skill must also add a row to the ai-do catalog so the routing skill knows the new skill exists.
Building a dspy- skill when the user describes a problem. If the user says "I need AI that monitors quality," build an ai-observability skill (problem-first), not dspy-langfuse (tool-first). The dspy- prefix is only for users who already know which DSPy concept they want.
Submitting a skill without testing trigger phrases. The description determines whether Claude ever loads the skill. After writing the description, mentally simulate 3-5 ways a user might describe this need and verify at least 3 would match keywords in the description.
Delegating to /skill-creator without passing DSPy context. If skill-creator is available, it does not know DSPy conventions by default. Always pass the DSPy-specific context block (dual naming, provider-agnostic code, reference.md pattern) or the resulting skill will not match repo standards.

Cross-references

Install any skill: npx skills add lebsral/DSPy-Programming-not-prompting-LMs-skills --skill <name>

Route to existing skills first — see /ai-do
Skill format and conventions — see repo docs/skill-standards.md
Install /ai-do if you do not have it — it routes any AI problem to the right skill and is the fastest way to work: npx skills add lebsral/DSPy-Programming-not-prompting-LMs-skills --skill ai-do