lebsral/ai-planning
skills/ai-planning/SKILL.md
Plan a multi-phase AI feature before building it. Use when you have a PRD or project idea and need to figure out the execution order, which skills to use in what sequence, or how to break an ambitious AI project into phases. Also use when you want to scope an AI feature, create a phased rollout plan, or figure out dependencies between AI components., help me figure out how to execute this, plan my AI feature, what order should I build this in, AI project roadmap, break this into phases, scope an AI feature, phased AI rollout, AI feature planning, multi-phase AI project, AI project dependencies, which skills do I need, AI execution plan
$ install --global
skillsauthnpx skillsauth add lebsral/dspy-programming-not-prompting-lms-skills ai-planningInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: May 5, 2026, 7:57 AM178.0s2 files scanned
SKILL.md
- name:
- ai-planning
- description:
- Plan a multi-phase AI feature before building it. Use when you have a PRD or project idea and need to figure out the execution order, which skills to use in what sequence, or how to break an ambitious AI project into phases. Also use when you want to scope an AI feature, create a phased rollout plan, or figure out dependencies between AI components., help me figure out how to execute this, plan my AI feature, what order should I build this in, AI project roadmap, break this into phases, scope an AI feature, phased AI rollout, AI feature planning, multi-phase AI project, AI project dependencies, which skills do I need, AI execution plan
Plan Your AI Feature
Use this skill when you have an AI project idea and need to figure out what to build, in what order, with which skills. ai-do routes to one skill per invocation. ai-kickoff scaffolds one project. Neither creates a multi-phase roadmap. This skill fills that gap.
Should You Even Plan?
Not every task needs a plan. Route directly if any of these match:
| Situation | What to do instead |
|---|---|
| Single task (one classification, one extraction) | Go straight to that skill |
| Already building something | Use the skill for what you are building |
| Just exploring or learning DSPy | Use /ai-kickoff |
| Fixing a broken feature | Use /ai-fixing-errors |
If none of those match, continue below.
Step 1: Answer 3 Questions
Before mapping skills or phases, answer these three questions. Write them down — they drive every decision in the plan.
Question 1: What is the end goal? What does "done" look like for users? Be specific. "AI-powered search" is not a goal. "Users type a question and get an answer with a citation to the source article within 2 seconds" is a goal.
Question 2: What do you have today?
- Existing code (prototype, integration, nothing)?
- Data (labeled examples, raw documents, a database, nothing)?
- Models already in use?
Question 3: What are the hard constraints?
- Timeline (launch in 2 weeks vs. 6 months)
- Budget (API cost limits, inference budget)
- Model restrictions (must use on-prem, no OpenAI, etc.)
- Compliance (no PII to external APIs, audit logging required)
Step 2: Map Capabilities to Skills
Find the capabilities your feature needs in the left column. The right column is the skill to use.
| Capability | Skill |
|---|---|
| Classify, sort, label, route | /ai-sorting |
| Extract structured data | /ai-parsing-data |
| Search documents, RAG | /ai-searching-docs |
| Answer database questions | /ai-querying-databases |
| Summarize content | /ai-summarizing |
| Generate text or content | /ai-writing-content |
| Take actions, call APIs | /ai-taking-actions |
| Multi-step reasoning | /ai-reasoning |
| Chain multiple AI steps | /ai-building-pipelines |
| Conversational AI | /ai-building-chatbots |
| Multiple agents | /ai-coordinating-agents |
| Score, grade, evaluate | /ai-scoring |
| Moderate content | /ai-moderating-content |
| Choose the right pattern | /ai-choosing-architecture |
| Review code quality | /ai-auditing-code |
Circle (or list) the 2-5 capabilities your project actually needs. If you circled more than 5, you are planning too much at once — pick the core path and defer the rest.
Step 3: Decide Phase Ordering
Match your situation to one of these starting points:
Greenfield — no code, no data
- Phase 1: Set up data collection + scaffold with
/ai-kickoff - Phase 2: Build the core feature skill
- Phase 3: Measure and improve with
/ai-improving-accuracy
Have data, no code
- Phase 1: Scaffold with
/ai-kickoff+ build core feature - Phase 2: Evaluate quality with
/ai-improving-accuracy - Phase 3: Productionize with
/ai-serving-apis
Working prototype, needs improvement
- Phase 1: Measure current quality with
/ai-improving-accuracy - Phase 2: Optimize (compile with an optimizer, tune prompts)
- Phase 3: Harden for production
Production system, adding new capability
- Phase 1: Build new capability using its skill in isolation
- Phase 2: Integrate into existing pipeline with
/ai-building-pipelines - Phase 3: Re-optimize the full pipeline
Step 4: Generate the Plan
Fill in this template. Keep Phase 1 detailed, Phase 2 directional, Phase 3 a placeholder until Phase 2 is done.
## AI Feature Plan: [Feature Name]
### Phase 1: [Phase Name] (start here)
- **Goal:** What this phase achieves
- **Skill:** `/ai-xxx` -- one sentence on why this skill
- **Deliverable:** What exists when this phase is done
### Phase 2: [Phase Name]
- **Goal:** ...
- **Skill:** `/ai-xxx` -- why
- **Deliverable:** ...
### Phase 3: [Phase Name]
- **Goal:** ...
- **Skill:** `/ai-xxx` -- why
- **Deliverable:** ...
### Dependencies
- Phase 2 needs Phase 1 output because [specific reason]
- Phase 3 needs Phase 2 output because [specific reason]
### What to skip for now
- [capability] -- reason it is not Phase 1 material
- [capability] -- reason it is not Phase 1 material
Common Project Archetypes
Five common project patterns with their natural phase sequence:
Support ticket triage system
Phase 1: /ai-sorting (classify + route tickets) -> Phase 2: /ai-summarizing (summarize for agents) -> Phase 3: /ai-building-pipelines (connect classify + summarize) -> Phase 4: /ai-improving-accuracy + /ai-serving-apis
Knowledge base or help center
Phase 1: /ai-searching-docs (index articles, answer questions with citations) -> Phase 2: /ai-stopping-hallucinations (add guardrails) -> Phase 3: /ai-improving-accuracy -> Phase 4: /ai-serving-apis
Document processing pipeline
Phase 1: /ai-parsing-data (extract fields from documents) -> Phase 2: /ai-checking-outputs (validate extracted fields) -> Phase 3: /ai-building-pipelines (batch processing) -> Phase 4: /ai-improving-accuracy + /ai-serving-apis
Content generation platform
Phase 1: /ai-writing-content (generate drafts at scale) -> Phase 2: /ai-scoring (grade quality before publishing) -> Phase 3: /ai-improving-accuracy (optimize for tone and brand) -> Phase 4: /ai-serving-apis
AI agent
Phase 1: /ai-choosing-architecture (pick the right pattern) + /ai-taking-actions (wire up tools) -> Phase 2: /ai-building-pipelines (multi-step coordination) -> Phase 3: /ai-testing-safety + /ai-monitoring
See examples.md for fully worked plans for each of these archetypes.
Gotchas
-
Planning all phases in equal detail. Only detail Phase 1 fully. Later phases will change once you see Phase 1 results. Detailed Phase 3 plans written before Phase 1 is done are fiction.
-
Recommending optimization before a baseline exists. Always build first, measure second, optimize third. Running a DSPy optimizer on a feature you have not yet evaluated is guesswork.
-
Skipping the data question. Most AI projects stall because there is no evaluation data. Surface this in Phase 1. If you do not know what good output looks like, you cannot tell when you are done.
-
Including production skills too early. Monitoring, tracing, and serving APIs are Phase 3 or later, not Phase 1. Do not let infrastructure planning crowd out feature planning.
-
Listing every possible skill instead of selecting 3-5. A plan that touches 12 skills is a catalog, not a plan. If your plan has more than 5 skills, cut the ones that are not on the critical path and put them in the "skip for now" section.
Cross-References
- For a specific skill recommendation on a single task, use
/ai-do - To scaffold a project once you know what to build, use
/ai-kickoff - To pick the right DSPy pattern for your use case, use
/ai-choosing-architecture - To review existing AI code for problems, use
/ai-auditing-code - Install
/ai-doif 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
Related Skills
lebsral/ai-watching-optimization
tools
VerifiedTrustedCommunity
See what is happening during optimizer.compile() instead of waiting blind. Use when you want to watch optimization progress, see scores as they come in, know if your optimizer is working, check if optimization is stuck, understand why optimization is taking too long, get live progress during compile, monitor convergence, detect overfitting during optimization, interpret optimization results, or pick the right tool for watching optimization. Also used for optimizer progress bar, is my optimizer doing anything, optimization seems stuck, how long will optimization take, watch GEPA run, watch MIPROv2 run, live optimization dashboard, optimizer not improving, scores not going up, optimization taking forever, see what optimizer is doing, debug slow optimization, optimization visibility, optimizer metrics, track compile progress, optimization observability.
6SKILL.mdUpdated May 31, 2026
lebsral/dspy-miprov2
testing
VerifiedTrustedCommunity
Use when you want the highest-quality prompt optimization DSPy offers — jointly optimizes instructions and few-shot demos, with auto=light/medium/heavy presets. Common scenarios - you want the best possible accuracy from prompt optimization, jointly tuning instructions and few-shot demonstrations, using auto presets for different compute budgets, or when COPRO or BootstrapFewShot alone are not reaching your accuracy target. Related - ai-improving-accuracy, dspy-copro, dspy-bootstrap-few-shot. Also used for dspy.MIPROv2, best DSPy optimizer, highest quality optimization, auto=light medium heavy, joint instruction and demo optimization, most powerful prompt optimizer, MIPROv2 vs COPRO vs BootstrapFewShot, which optimizer should I use, state of the art prompt optimization, when to use MIPROv2, optimize both instructions and examples, heavy optimization for production, best optimizer for accuracy.
6SKILL.mdUpdated Apr 27, 2026
lebsral/dspy-langwatch
testing
VerifiedTrustedCommunity
Use LangWatch for DSPy auto-tracing and real-time optimizer progress. Use when you want to set up LangWatch, langwatch.dspy.init, auto-tracing DSPy, real-time optimization dashboard, optimizer progress tracking, app.langwatch.ai, or DSPy optimizer dashboard. Also used for langwatch setup, pip install langwatch, langwatch trace, optimizer progress, real-time optimization, watch optimizer run, LangWatch self-hosted, langwatch docker, langwatch vs langtrace, langwatch autotrack_dspy.
6SKILL.mdUpdated Apr 27, 2026
lebsral/dspy-gepa
data-ai
VerifiedTrustedCommunity
Use when you want to optimize instructions without few-shot examples — a lightweight alternative to COPRO when you do not have or do not want to use demonstrations. Common scenarios - optimizing instructions when you do not have or do not want to use few-shot demonstrations, lightweight instruction search as a first step, tasks where examples in the prompt confuse the model, or when you want fast instruction optimization without the cost of COPRO. Related - ai-improving-accuracy, dspy-copro, dspy-miprov2. Also used for dspy.GEPA, instruction optimization without demos, lightweight prompt optimization, optimize instructions only, no few-shot examples needed, GEPA vs COPRO, quick instruction search, when demonstrations hurt performance, zero-shot optimization, instruction-only optimizer, simplest instruction tuner, fast prompt optimization, skip few-shot and just tune instructions, optimize Pydantic field descriptions, GEPA structured output, GEPA does not optimize field desc.
6SKILL.mdUpdated Apr 27, 2026