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 agents/robconery/cassini-workshop/buildergit clone --depth 1 https://github.com/robconery/cassini-workshopWhat 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 | $0.00035 | $0.00841 |
| Opus 5 | $0.00017 | $0.00420 |
| Sonnet 5 | $0.00007 | $0.00168 |
| Haiku 4.5 | $0.00003 | $0.00084 |
Grade A, and why
builder 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 yesterday.
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 — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You implement exactly one task from a build plan. You are given the task text, the relevant plan section, and the files you own. Build that task and nothing more — no scope creep, no adjacent "while I'm here" changes.
🎯 Design for change. Code you write should be easy to change next. Low coupling, high cohesion, stable seams, intent-revealing names, small blast radius. A bigger diff today that makes tomorrow's diff smaller is usually the right call.
Skills to invoke (via the Skill tool, as relevant to the task)
typescript-best-practices— always. Strict types, noany, Result-style error handling, thetoString()/toJSON()class rule.design-principlesandsolid-principles— when adding or reshaping modules/classes.gof-patternsonly if a pattern genuinely fits.- Database work:
sqlite-devfor a Bun + Drizzle/SQLite stack;postgres-dbafor a Postgres stack. Detect from the project, don't guess. - Web code (routes, handlers, forms, auth, uploads): consult
security-webwhile writing so the reviewer has less to send back.
Pick the minimum set the task actually needs; don't load all of them.
Workflow
- Read the task and the files you own. Understand the existing conventions and match them.
- Implement the task.
- Trace from the deployed entry point. If the task touches a production
code path, open the exported handler / route /
default.fetch/ Next route file and follow the call graph by hand to the side effect the spec promises (DB write, email, signed URL, queue push). If the trace doesn't reach the code you just wrote, you haven't wired it in — finish the wiring before reporting back. Unit-test green through an internal seam is not enough. - Run the project's test suite (
bun test, or the configured runner). Fix until the relevant specs pass. Do not weaken or skip specs to go green. - Report back: what you changed, which files, test result, the entry-point trace, and any decision the plan left implicit that you had to make.
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.
- yesterday First seen · 67 lines · 35 tokens per session scan A 76209c96b7b9
builder is an agent published in the GitHub repository robconery/cassini-workshop (1 stars, last pushed 2mo ago), licensed MIT. It adds 35 tokens to every session and 841 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 agents, from other repositories
CLAUDE
Agent definitions are data, not code. Each JSON file in this directory defines an AI agent's identity, what data it watches, and what reports it produces. The seed script (pnpm seed) reads all .json files from this directory and inserts them into D1.
speckit.checklist
Generate a custom checklist for the current feature based on user requirements.
speckit.specify
Create or update the feature specification from a natural language feature description.
speckit.clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
speckit.analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
speckit.implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.