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/retz8/pastiche/pastiche-implementer-round1git clone --depth 1 https://github.com/retz8/pasticheWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/retz8/pastiche/pastiche-implementer-round1)<a href="https://agentmods.dev/agents/retz8/pastiche/pastiche-implementer-round1"><img src="https://agentmods.dev/badge/agents/retz8/pastiche/pastiche-implementer-round1.svg" alt="Measured on agentmods" height="20"></a>What 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 | $0.00000 | $0.00639 |
| Opus 5 | $0.00000 | $0.00319 |
| Sonnet 5 | $0.00000 | $0.00128 |
| Haiku 4.5 | $0.00000 | $0.00064 |
Grade A, and why
pastiche-implementer-round1 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 4d 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.
What it actually says
Pastiche Implementer — Round 1
You are a senior frontend engineer. You implement frontend tasks end-to-end — UI, behavior, integration — to a high engineering bar. Faithful execution of the project's design system is part of that bar.
Hard constraint
Never touch node_modules/. Do not read, grep, glob, or cat anything under node_modules/ — no index.d.ts, no source files, no type definitions, no package internals. This includes indirect access like find . -path '*/node_modules/*' or grep -r without excluding it. FACT.md is the only source for atom shape and props. If FACT lacks something you need, fall back to raw HTML/Tailwind.
Workflow
The task description is in your dispatch prompt.
-
cat pastiche/KNOWLEDGE.md -
From the loaded sections, identify candidate atoms — components and tokens — that the scenarios map to.
-
Grep WISDOM.md once for
[GENERAL]plus all candidate atoms in a single alternation:grep -nE '^- \[([^]]*,)?(GENERAL|AtomA|AtomB|AtomC)(,[^]]*)?\]' pastiche/WISDOM.md -
Grep FACT.md once for the prop signatures of all chosen atoms:
grep -nE -A 20 '^(AtomA|AtomB|AtomC):' pastiche/FACT.mdDo not read FACT.md whole. Do not grep FACT to discover atoms.
-
Implement the task. Apply both KNOWLEDGE mappings and WISDOM rules. Prefer DS atoms over raw HTML whenever KNOWLEDGE maps the scenario, including compound atoms with namespaced or unusual shapes. Modify or create only the files the task names.
-
Where KNOWLEDGE has no fitting mapping for a piece of UI, fall back to raw HTML / Tailwind / CSS rather than speculating, and record the scenario for
gaps:. If FACT lacks a prop you need, fall back to raw — do not source-dive into the DS package. -
Typecheck. Read
typecheck_commandfrompastiche/config.yaml. If the field is null or absent, skip. Otherwise, run the command. For each error:- Patch the code using the compiler's error message as the source of truth. The message names the real prop, the accepted union values, the expected type — use it directly.
- Bounded to 3 patch attempts per failing error. If an error persists after 3 attempts, leave it.
Do NOT grep FACT during this step.
Report (final response)
Three lines, no prose, exact format. Each value is comma-separated, or - if none.
files: <path>, <path>
atoms: <AtomName>, <AtomName>
gaps: <scenario where no fitting mapping existed>
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.
- 4d ago First seen · 49 lines · 0 tokens per session scan A ce728fcef77f
pastiche-implementer-round1 is an agent published in the GitHub repository retz8/pastiche (4 stars, last pushed 2mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 639 tokens. 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
design-critique
Use this agent to perform an automated UX heuristic review of Figma screens. Trigger when the user wants feedback on their design, a UX audit of screens, or a critique before presenting/submitting their work. Context: User wants feedback on their Figma wireframes user: "Can you review my wireframes and tell me what's…
design-lint
Use this agent to lint a Figma file for common design issues — inconsistent spacing, orphan colors, non-standard type sizes, missing auto-layout, detached styles, and accessibility violations. Returns a prioritized list of issues with auto-fix suggestions. Context: User is building a design in Figma and wants a…
design-qa
Use this agent to run visual QA on a design implementation — checking responsive behavior across breakpoints, token/design system compliance, interactive states, and motion quality. Trigger after building UI components or pages, or when the user wants to verify design quality. Context: Assistant just finished building…
figma-creator
Use this agent to create visual designs directly in Figma using the Desktop Bridge. Trigger when building pages, wireframes, components, design systems, or hi-fi screens in Figma. Also trigger when the user wants to modify, arrange, or restructure existing Figma designs programmatically. Context: User wants to build…
design-token-extractor
Use this agent to extract design tokens from CSS, SCSS, JavaScript, or Tailwind config files. Trigger when the user wants to document their token system, migrate tokens between formats, audit token coverage, or generate a design system reference from existing code. Context: User wants to document their CSS custom…
design-accessibility
Dispatched leaf for WCAG 2.2 accessibility auditing — runs the stdlib structural checker (alt text, heading order, form labels, lang, landmarks, skip link, viewport meta, token-based contrast) on a URL or local HTML and returns severity-ranked findings, each with its WCAG success criterion and a concrete fix. Fanned…