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 skills add jcooley8/pincushion-plugin --skill pincushiongit clone --depth 1 https://github.com/jcooley8/pincushion-pluginWrote 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/skills/jcooley8/pincushion-plugin/pincushion)<a href="https://agentmods.dev/skills/jcooley8/pincushion-plugin/pincushion"><img src="https://agentmods.dev/badge/skills/jcooley8/pincushion-plugin/pincushion/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/jcooley8/pincushion-plugin/pincushion"><img src="https://agentmods.dev/badge/skills/jcooley8/pincushion-plugin/pincushion.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00039 | $0.00842 |
| Opus 5.5 | $0.00016 | $0.00337 |
| Sonnet 5.5 | $0.00008 | $0.00168 |
| Haiku 4.5 | $0.00004 | $0.00084 |
Grade A, and why
pincushion 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 7d 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
Connect Pincushion
- Confirm the current working directory is the user's app repository, not this
plugin's installation directory. Discover the
pincushionMCP tools with Grok's tool search. Readget_project_contextand, when identity is unclear,get_project_identity_diagnostic. These are read-only; never register a project just to look it up. If the MCP is bound to another directory, stop and restart Grok from the intended repository before writing any pins. - If connection is required, use the existing browser sign-in:
npx --yes [email protected] login. Let the user complete sign-in; never ask for a license key, read credential files, or copy credentials into chat, source, plugin configuration or logs. Refresh/mcpsafterward. An absent credential is a connection blocker, not an empty project. - Obtain the intended project name, exact page URL, existing project ID when present, and desired comment access. If several projects or URLs could match, ask the user to choose; never select the first result. Preserve an existing project's identity and access mode. For a new project explain open, domain-restricted and invite-only access using the server's current plan capabilities; do not silently make a private app open.
- Read the app's own brand documents, README, representative routes and theme tokens. Draft a concise brand brief covering audience, purpose, tone, visual identity and exclusions. Ask only for missing context. Do not inspect .env, credentials, unrelated repositories or private customer records. If the brief cannot be grounded, stop before critique rather than inventing it.
- Explain the first Crit before asking for confirmation: project URLs and the approved brand brief are stored by Pincushion; Crit aims for at least three distinct useful AI findings when the rendered evidence supports them, with at most three per page. Findings with selectors and relevant element context are synced as pins accessible under the project's access settings. Grok processes the sources and screenshots it reads through the user's model session. Local screenshots stay local unless the user also asks for a share report. A share report uploads selected screenshots/positions and creates a native link. Public reports are accessible to anyone with the link; private reports require authenticated project membership. Private/authenticated screens require deliberate sharing approval and a member-only report.
- Once the user confirms the project, access and first Crit, call
configure_projectwith that exact identity, URLs, approvedcritiqueContext(maximum 8192 characters), relevantcritiquePolicy, andautoCritique: false. Do not set deploymentAutomation or add members, hooks, source injection or subscriptions. Inspect the response for errors, then read backget_project_context({ projectId })and verify identity, URLs and context. Do not proceed through connection_required, an access error or a mismatch. - Run
/pincushion-critimmediately for the confirmed URL and project. Do not silently stop after registration. If a prerequisite blocks the Crit, state the precise blocker and resume after it is resolved. A setup completion is not a Crit completion. Follow the Crit skill's zero-finding behavior. - Offer
/pincushion-pinsto review results and/pincushion-implementonly when the user wants approved findings fixed. Never approve or implement findings automatically. After later completed UI work, suggest another Crit once per batch using the checkpoint skill.
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.
- 7d ago Changed · +2 lines d2dc88d2e2a8
- 8d ago First seen · 57 lines · 39 tokens per session scan A 5224498863b8
pincushion is a skill published in the GitHub repository jcooley8/pincushion-plugin (1 stars, last pushed 6d ago), licensed MIT. It adds 39 tokens to every session and 842 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-09-28.
Other skills, from other repositories
webgl-holographic-foil
A self-contained WebGL2 hero: thin-film interference over a crushed-foil surface whose palette shifts with the viewing angle; move the cursor to tilt the film.
html-ppt-hermes-cyber-terminal
OpenDesign + BYOK: choosing and wiring your own model, hands-on — cost, quality, and the routing decision. Built as a decision-grade AI literacy deck for engineers, IT, applied-AI teams.
html-ppt-taste-brutalist
16:9 HTML deck in tactical-telemetry / CRT-terminal taste. Deactivated-CRT charcoal slides, white-phosphor monospace, hazard-red accent, scanline overlay, ASCII syntax, density over decoration. Distilled from Leonxlnx/taste-skill brutalist-skill (Tactical Telemetry mode).
visual-ralph
Visual Ralph orchestration for frontend UI from generated references, static references, or live URL targets, using $ultragoal with built-in visual verdict and pixel-diff evidence until the implementation matches and leaves a reproducible design system.
accessibility
Consolidated accessibility skill entrypoint for WCAG 2.2, ARIA Authoring Practices, cognitive accessibility, Section 508, EN 301 549, design intent verification, and the Accessibility Planner workflow.
build-ui-components-and-systems
A guide for building reusable UI components and design systems: shared visual rules, component interfaces, behavior, accessibility requirements, documentation, packaging, and version management. A design system is a maintained set of parts and rules that multiple screens or products can share.