MCP server that scans web pages for WCAG accessibility issues with axe-core, so AI agents can audit and fix accessibility. By
Free
Nothing to run from here
Pincushion is installed from its publisher's own source and answers where it runs, so this marketplace is not in the path of a single call. There is no address here to send one to, and a panel that pretended otherwise would be showing you an answer we made up. Install it and call it from your own client — the Installation tab has the entry for each one.
Open Installation—
p95 latency
—
call success
local
runs on your machine
—
calls last 7d
What it does
Visual feedback as agent work packets. Stakeholders drop pins on your live app; your AI coding agent reads each pin (selector, screenshot, DOM, thread, acceptance criteria) via MCP and ships the fix.
Quickstart
# 1 — run it from where its publisher ships it
npx -y pincushion-mcp
# 2 — ask your agent something
> Visual feedback as agent work packets. Stakeholders drop pins on your live app; your AI coding agent reads each pin (selector, screenshot
Pincushion is free: there is no plan to choose, no cap to set and nothing that can bill you.
Publishernot claimed
JCAll 1 server
jcooley81 server
1 category
Collected from a public index. Nobody has claimed this account, so nothing here was written by its author — claim it if it is yours.
Installation
Where are you running it?
Every route below installs the same thing and ends at the same approval screen. This one runs on your machine: your client starts Pincushion as a process under your own user, with your files and your network, so the tool surface below is what it can do to your computer rather than to a server somewhere else. It is scanned, signed and pinned to the version you choose — read the surface before you approve it.
Claude Codedeclared by this listing
This is a public server: you run it yourself and this marketplace is not in the path. Claude Code registers it in one command.
claude mcp add pincushion-mcp -- npx -y pincushion-mcpReconnect, or start a new session, and the tools appear in the model’s tool list.
Free to install, and it runs from its publisher’s own package: no account, no key, no card.
What it writes
One config entry your client uses to start the process locally. A local server runs with your file system and your network, which is why it is priced without metering.
What you approve
37 tools, with what each one reads, writes and reaches shown before you agree — the same list on every route above. Read the tool surface.
Tool surface
What the model actually sees. Descriptions are diffed on every release — see version history.
Retrieve annotation pins from the .feedback/ directory. Filter by page URL, LWC component name, or status. Use this to understand what feedback exists before making changes.
Takes no parameters.
Example callreturns →
Full-text search across all annotation comments, selectors, component names, and tags.
Takes no parameters.
Example callreturns →
Get a high-level rollup of all open feedback: counts by status, page, and component. Use this to plan what to address first.
Takes no parameters.
Example callreturns →
Get all feedback pins targeting a specific LWC component, with a plain-language summary ready for implementation. Returns element selectors, comments, and thread history.
Takes no parameters.
Example callreturns →
Mark an annotation as resolved after addressing the feedback. Optionally add a resolution comment explaining what was changed.
Takes no parameters.
Example callreturns →
Add a reply to an annotation thread (e.g. to ask a clarifying question or note a finding).
Takes no parameters.
Example callreturns →
Used by /pincushion-replies. Returns pins where Pincushion AI should respond, with each candidate tagged by trigger reason. Two triggers: (a)
Takes no parameters.
Example callreturns →
Post a Pincushion AI reply to a pin\
Takes no parameters.
Example callreturns →
Lists production deployment frame-refresh work independently of the Critique queue. Use this even when autoCritique is false. The result is redacted: it contains no deployment URL/hash, browser state, storage path, or capture proof. Start a selected run before capturing.
Takes no parameters.
Example callreturns →
Requests server-side frame-refresh completion after all exact page/device receipts have been recorded. The server rejects missing, stale, replayed, cross-deployment, or expired capture-run evidence and never treats a queue state as a current frame.
Takes no parameters.
Example callreturns →
Mark a critique_queue request as completed after the critic subagent has run on its page URLs. Pass the request id (from get_pending_critiques) and the total pin_count produced. The server scopes the update to your license — you cannot complete another tenant\
Takes no parameters.
Example callreturns →
Register the exact bounded URL inventory discovered in one local browser context before staging deployment captures. The backend returns accepted pending pages; upload snapshots and record terminal results only for that returned subset, never for a locally guessed URL.
Takes no parameters.
Example callreturns →
Resolve a pin after applying a code fix. Transitions the pin directly to
Takes no parameters.
Example callreturns →
Compute median + p25/p75 time-to-fix from resolved pins. Returns sample size + threshold flag so callers can honestly hide the metric when the dataset is too small (< 5 resolved pins). This is the marketing proof point that distinguishes Pincushion from
Takes no parameters.
Example callreturns →
Record the outcome of Pincushion AI\
Takes no parameters.
Example callreturns →
Link a deploy URL to a resolved pin. Typically called by the deploy-hook edge function once a deploy that includes the pin\
Takes no parameters.
Example callreturns →
Get instructions for setting up and connecting the Pincushion browser extension to this MCP server. Includes configuration examples for Cursor, Claude Desktop, Claude Code, VS Code, Windsurf, and other agents.
Takes no parameters.
Example callreturns →
Finish the quickstart demo loop after editing the demo copy. Confirms the change, returns the bridge to real usage, and cleans up the throwaway .feedback/.quickstart/ files. Does NOT create or resolve any real pin.
Takes no parameters.
Example callreturns →
Read-only lookup of a project\
Takes no parameters.
Example callreturns →
Register a Pincushion project and associate it with your app\
Takes no parameters.
Example callreturns →
Lightweight write-only path for the layered critique-context system. Use this from /setup and /refresh-brand after the dev agent has gathered repo signals (README, theme tokens, sample copy, competitor URLs, recent resolved pins) and compiled them into a critique brief. Unlike
Takes no parameters.
Example callreturns →
Get all pins waiting for developer attention. Returns three categories: (1)
Takes no parameters.
Example callreturns →
Claim an actionable pin before starting work on it. Transitions the pin from
Takes no parameters.
Example callreturns →
Mark a pin as approved for implementation. Only approved pins should be implemented by agents. This transitions the pin from
Takes no parameters.
Example callreturns →
Assign a pin directly to your local coding agent. Promotes the pin to
Takes no parameters.
Example callreturns →
Get a single implementation packet for one page URL. Useful when you want to batch-fix one page at a time. Returns the same shape as a single entry in implement_approved_pins.packets — pins, aggregated selectors, suggested branch, traceability config. Matches by exact URL or partial substring.
Takes no parameters.
Example callreturns →
Get pins that the developer has selected for implementation from the dashboard or PINS.md checkboxes. Returns the selected pin IDs with full context (element, thread, deep link). Use this to know which pins the developer wants you to work on next.
Takes no parameters.
Example callreturns →
Add a collaborator to a Pincushion project. Developers consume a paid seat and can implement pins. Commenters are free and unlimited. Returns an upgrade prompt if the seat limit is reached.
Takes no parameters.
Example callreturns →
List all members of a Pincushion project with their roles, plus seat usage info.
Takes no parameters.
Example callreturns →
Remove a collaborator from a Pincushion project. Frees up the seat if they were an editor.
Takes no parameters.
Example callreturns →
Turns supplied real captures into a report receipt; it does not inspect a page or generate feedback. By default it mints a Crit and therefore requires the exact Pincushion AI pin IDs the critic just created, each positioned in a supplied capture. Set purpose:
Takes no parameters.
Example callreturns →
Generate an Add-to-Slack OAuth URL pre-bound to a project. Most users should prefer the public storefront URL (also returned, https://pincushion.io/install/slack) — since May 2026, that auto-links to a Pincushion license when the installer\
Takes no parameters.
Example callreturns →
List Slack and Microsoft Teams webhook subscriptions for a Pincushion project. Webhook URLs are masked.
Takes no parameters.
Example callreturns →
LEGACY FALLBACK. Since May 2026, Slack installs auto-link to a Pincushion license when the installer\
Takes no parameters.
Example callreturns →
Remove a Slack, Microsoft Teams, or Discord webhook subscription from a Pincushion project.
Takes no parameters.
Example callreturns →
Preview the Slack, Teams, or Discord notification shape and recommended event routing before connecting a real webhook.
Takes no parameters.
Example callreturns →
Read or update the caller\
Takes no parameters.
Example callreturns →
Freeyour plan
$0 forever
- Every tool, no call limit
- No card, no account needed
- Source published under a licence you can read
- Runs on your machine — nothing of it reaches our gateway
- Nothing to cap, because nothing bills
What counts against your monthly calls
| Tool | Unit | Calls used | Out of the allowance |
|---|
Nothing here is billable. Pincushion costs nothing to install and nothing to call, at any volume.
Call allowance · this installenforced at the gateway
Monthly allowance, in calls
—
—
Alert at 80% · then the tool returns a normal MCP error
Used this month: $8.42
$16.58 left
ATrust score 93 / 100
Release 1.11.12 · scanned 25 Aug 2026
Two independent axes, because powerful and malicious are different questions. The grade is threat only. The capability level is blast radius, and it is never a penalty on the grade — it is priced as one subtract-only term in the score, where you can see it.
Threat score100
Client score93
Findings1 threat · 10 capability
Capabilityhigh
Coveragesource
Verificationrepo
Surface digest1b315e8ab41070766915b702e8fbfaf6348c34a56a6ebd639f358bba1ce7624b
Score vectorthreat 100 → client 93
| Term | Level | What it prices | Points |
|---|---|---|---|
| capability-exposure | high | capability blast radius (high) — client exposure if the model is manipulated | −6 |
| verification-discount | repo | publisher verification (public source) — no provenance, but the source is public and inspectable | −1 |
| coverage-honesty | source | inspection depth (source) — how much of the target the scan could see | −0 |
score = clamp(round(threatScore − E_cap − E_ver − E_cov), 0, 100). No model, no opinion, no network call at scoring time: the same methodology version against the same bytes always produces this number, and every line above is a field of the report rather than a summary of it.
What the scan could actually read
A grade is only as meaningful as its coverage, so the scanner publishes its own depth before it publishes its result.
Tool surfaceread
Implementation sourceread
Package metadataread
Live transportnot available
Tools were statically extracted from the published source (66 recovered), not enumerated from a running server. Tool-poisoning, Unicode-smuggling, capability and toxic-flow analysis ran on this inferred surface, but a mis-parsed registration could be missed or mis-attributed, so tool-derived findings are capped below “confirmed”. To grade the real runtime surface, scan the running server: --command "npx -y <package>".
Capability — what it could do if the model were manipulated
Tags derived from each tool’s schema and the implementation, not from what the tool calls itself. high is the level these add up to.
| Tool | Capability tags | Why the tag was assigned |
|---|---|---|
| Verification | no tags | |
| get_annotations | no tags | |
| search_annotations | no tags | |
| get_feedback_summary | no tags | |
| get_component_feedback | no tags | |
| resolve_annotation | no tags | |
| add_agent_reply | no tags | |
| get_reply_candidates | untrusted-input | |
| add_bot_reply | untrusted-input | |
| get_pending_critiques | no tags | |
| get_pending_deployment_frame_refreshes | no tags | |
| start_deployment_capture_run | no tags | |
| complete_deployment_capture_run | no tags | |
| complete_critique_request | no tags | |
| register_deployment_discovered_pages | no tags | |
| record_deployment_page_result | no tags | |
| create_critique_pin | no tags | |
| create_agent_pin | no tags | |
| fix_and_resolve | untrusted-input | |
| get_time_to_fix_metrics | no tags | |
| record_pin_verification | no tags | |
| link_pin_deploy | no tags | |
| get_setup_instructions | no tags | |
| start_quickstart_demo | no tags | |
| resolve_quickstart_demo | no tags | |
| get_project_context | no tags | |
| configure_project | no tags | |
| get_deployment_reconciliation_status | no tags | |
| manage_deployment_automation | external-sink | |
| update_critique_context | no tags | |
| get_actionable_pins | no tags | |
| claim_pin | no tags | |
| approve_pin | no tags | |
| assign_pin_to_agent | no tags | |
| implement_approved_pins | no tags | |
| get_implementation_packet | no tags | |
| get_selected_pins | no tags | |
| add_member | no tags | |
| list_members | no tags | |
| remove_member | no tags | |
| create_invite_link | no tags | |
| create_share_report | no tags | |
| upload_page_snapshot | no tags | |
| generate_critique_report | no tags | |
| configure_collaboration_integration | external-sink | |
| create_slack_install_link | no tags | |
| list_collaboration_integrations | no tags | |
| claim_pending_slack_install | external-sink | |
| remove_collaboration_integration | external-sink | |
| preview_collaboration_notification | external-sink | |
| set_slack_preferences | no tags | |
| pins | no tags | |
| my-pins | no tags | |
| username | no tags | |
| resolve | no tags | |
| pin_id | no tags | |
| feedback-summary | no tags | |
| setup | no tags | |
| urls | no tags | |
| implement | no tags | |
| invite | no tags | |
| emails | no tags | |
| project_id | no tags | |
| report_url | no tags | |
| page_url | no tags | |
| pincushion-quickstart | no tags |
Toxic-flow graph
The lethal trifecta, checked as a graph rather than as a checklist: untrusted input, a sensitive source and an external sink have to meet before there is a path worth worrying about.
Untrusted input can drive an external action even though no sensitive source is exposed.
Supply chain and provenance
Integrityfirst-seen
Surface digest1b315e8ab4107076
Scanned2026-08-25
This is the first scan of this surface here, so there is nothing yet to compare it against.
Every result on this tab comes from one deterministic pass over the published package — offline, rule by rule, and auditable line by line above. Same methodology version, same bytes, same score.
Release history
Pinned to 1.11.12 — the install command below asks for that release. A pin is part of an install, so it is kept for this visit and written down when you install.
1.11.12
undated
installedNo release note was published with this version.
—
no reviews yet
Reviews are earned
Only accounts with at least 50 real tool calls against this server in the last 90 days can post. Ratings are weighted by how much the reviewer actually uses it, and publishers can reply once per review.
Reviews are earned
Writing one takes an account with at least 50 real tool calls against Pincushion in the last 90 days. That is the whole gate — there is no other way to post, which is why the counts beside each review are worth reading.
No reviews yet
Nobody has reviewed this listing. The rating on the card is the mean of the reviews written here and nothing else, so there is no rating until somebody writes the first — which takes an account with 50 real tool calls against it.