Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/stellarshenson/claude-code-plugins/program-writernpx skills add stellarshenson/claude-code-plugins --skill program-writergit clone --depth 1 https://github.com/stellarshenson/claude-code-pluginsWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00042 | $0.01073 |
| Opus 5 | $0.00021 | $0.00536 |
| Sonnet 5 | $0.00008 | $0.00215 |
| Haiku 4.5 | $0.00004 | $0.00107 |
Grade A, and why
program-writer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Program Writer
Write PROGRAM.md via dialogue. NEVER produce full document on first attempt. Build incrementally via questions and proposals.
Process
Round 1: Extract the intention
Read seed prompt (text after /autobuild). Goal: understand INTENTION (what + why), not HOW. ASK all in ONE message:
- End state? What does "done" look like? Outcome, not implementation
- What exists today? Current state. Gap to end state?
- Why does this matter? Problem solved? Cost of not doing? Shapes priority
- Off-limits? Files, behaviors, APIs that MUST NOT change
- How do we know it works? Test suite, benchmark, manual check?
- Biggest risk? What could waste iterations?
Users know WHAT but describe HOW. Separate intention from implementation. Follow up when unclear: "You mentioned refactoring X - goal = reduce complexity, improve testability, or enable a new feature?"
Do NOT proceed until intention crystal clear.
Round 2: Propose the program
Write first draft with:
- Objective (1-3 sentences, measurable)
- Current State (what exists, broken, baseline numbers)
- Work Items (flat list: scope, acceptance, priority)
- Exit Conditions (see below)
- Constraints (what not to change)
ASK about exit conditions: "When should orchestrator stop?
- Score stagnation (recommended) - stop if no improvement 2 consecutive iterations
- Score target - stop when score hits value (e.g. score < 5)
- Scope completion - stop when all work items meet acceptance
- Combined - whichever first: target OR stagnation OR scope complete
Specific, or default (stagnation + scope completion)?"
Present full program. Ask: "What's missing, wrong, over-scoped?"
Round 3+: Refine
User may:
- Add forgotten work items
- Remove out-of-scope items
- Change priorities
- Tighten acceptance
- Add constraints
Each round: update, show diff, ask if ready.
Final: User approval
Explicit approval only. Phrases: "looks good", "approved", "let's go", "run it", "start", "yes".
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 108 lines · 42 tokens per session scan A bbc509ff5c25
program-writer is a skill published in the GitHub repository stellarshenson/claude-code-plugins (3 stars, last pushed 2d ago), licensed MIT. It adds 42 tokens to every session and 1,073 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
auto-perf-optimize
Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.
chat-perf
Run chat perf benchmarks and memory leak checks against the local dev build or any published VS Code version. Use when investigating chat rendering regressions, validating perf-sensitive changes to chat UI, or checking for memory leaks in the chat response pipeline.
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…