workday-mcp: Instructions file for Claude Code

CLAUDE.md

workday-mcp CLAUDE.md is an instructions file for Claude Code from chrischall/workday-mcp. It costs 2,156 tokens per session, scanned A, original, MIT.

Project-specific instructions for contributing to workday-mcp, a read-only server that retrieves and interprets Workday data. They describe its architecture, coding conventions, security rules, and pull-request practices.

In plain words
What is it for?
Use them when changing the parser, redaction logic, Workday page handling, request URLs, tests, pull requests, or release notes.
Why use it?
They give an agent the project knowledge needed to make compatible changes and avoid exposing session tokens or other sensitive data.

Instructions file for Claude Code

Written for Claude Code: the file is CLAUDE.md. Also seen: reads .claude/ paths; mentions CLAUDE.md.

This is chrischall/workday-mcp's own configuration. It tells Claude Code how to work on workday-mcp itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything workday-mcp configures →

Reuse

Borrowing it

Nothing to install: this file belongs to chrischall/workday-mcp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/chrischall/workday-mcp/main/CLAUDE.md
Clone the repo
git clone --depth 1 https://github.com/chrischall/workday-mcp

Made for: Claude Code.

Wrote 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.

agentmods badge for workday-mcp CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/chrischall/workday-mcp/claude-md/github.svg)](https://agentmods.dev/instructions/chrischall/workday-mcp/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/chrischall/workday-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/chrischall/workday-mcp/claude-md/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.

agentmods 80×15 button for workday-mcp CLAUDE.md

Your own site · 80×15
<a href="https://agentmods.dev/instructions/chrischall/workday-mcp/claude-md"><img src="https://agentmods.dev/badge/instructions/chrischall/workday-mcp/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 2,156 This file is loaded in full into every session.
When invoked 2,156 The same file — it is already loaded in full.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.02156 $0.02156
Opus 5.5 $0.00862 $0.00862
Sonnet 5.5 $0.00431 $0.00431
Haiku 4.5 $0.00216 $0.00216

Measured 5d ago against content hash 0d7b4bfb765e, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

workday-mcp CLAUDE.md 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 5d 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.

CLAUDE.md · 126 lines

How it starts

The opening of the file, as written. The whole thing — 126 lines — stays where its author put it; the contents beside it link to each section on GitHub.

workday-mcp — contributor notes

Read-only MCP server for Workday, on the chrischall fleet skeleton (@chrischall/mcp-utils + @fetchproxy/server). Full-fetchproxy archetype: every request rides the user's signed-in *.myworkday.com tab.

Architecture

  • src/parse.ts — the durable core. Parses Workday's *.htmld widget-tree JSON into a flat, secret-free WorkdayTask, tagged with a kind. Handles all seven page families: data cards (flat + list rows), grid tables, fieldSet/panel detail pages, compositeView worker profiles, hierarchyNavigator org charts, landingPage hubs, and report prompt forms. Row objects are keyed by clean column names (label/value) whose propertyName is namespaced like wd:Label — key on the object key, NOT propertyName. Uri templates come in three dialects ({id}, [IID], already-concrete) and a moniker.target (URL-encoded absolute) beats a template. Most uris need .htmld appended or they 404. Never emits sessionSecureToken or other envelope secrets — it reads an explicit allowlist of fields only.
  • src/redact.ts — the DUAL of the parser's allowlist: a denylist applied to raw envelopes returned by workday_fetch / workday_graphql, covering secret and PII key names — including identity-document collections such as driversLicenses / identifications / visas / customIds (GraphQL's field-keyed JSON has no labels) — plus values whose sibling label (a string, or a list-card row's label column widget) or grid column names government/financial PII. Key-name matching is only sound because assertReadOnlyGraphql refuses aliasing such a field. getTask runs it BEFORE parseTask too, because the allowlist keeps envelope secrets out but copies widget values verbatim.
  • src/client.ts — deferred config (WORKDAY_TENANT required, WORKDAY_HOST defaulted), the fetchJson/postJson primitives, path normalization (strips #fragments, converts copied /d/ SPA URLs to data endpoints, appends .htmld), error mapping (non-2xx, SSO/sign-in bounce → SessionNotAuthenticatedError), first-call patience (send: until the bridge has answered once in this process, a timed-out GET is retried — the extension can take 2–60+ s to connect; POSTs are never retried), and the composed reads: crawl (hub → child cards, bounded by depth/maxCards/visited), resolveApp/openApp (by LABEL, never a captured id), getWorker/getWorkerTask/getOrgChart, getTeam/getReviews/getReview/getFeedback, getRequestForm, getInbox (a JSON API that wants the token from boot-config's appRoot, NOT the envelope token), getMyRequests, getTeamSnapshot, the request reads/writes (listOpenRequests/getRequest/planRequestUpdate/ applyRequestUpdate), and a query-only graphql. getRequestForm is a non-GET read (runReport is the other): Workday only renders a request's form inside the Create Request flow, so it drives that flow with form-urlencoded postForm steps and ALWAYS cancels it. _eventId_submit there is the PROMPT page's OK — the same event on a loaded form submits it, so it is refused once a page carries the bpfButtonBar. Never add a step that names a bpfButton. Tenant-specific ids are learned at runtime — the worker-profile uri prefix comes from a live org-chart node, never a constant.
  • src/parse-benefits.ts — the benefits pages (parse.ts' allowlist rule applies): your related-task catalog, an enrollment event's window, the elections landing page, a plan type's chunkedEditGrid of plans on offer, a plan's rate table, and elections grids → typed elections. Reads command buttons only when they are GETs — the enrollment pages carry POST buttons (Save for Later, Review and Sign, Confirm and Continue) that change elections. getBenefits / getBenefitEnrollment in client.ts are GET-only; tools/benefits.ts registers them.
  • src/parse-reports.ts — report prompt forms (prompts + current values, validation errors by prompt) and dashboards (tabs, worklets → workletIid, GET menu links). runReport in client.ts is the second non-GET read: it drives a prompt report's own flow (remove/search/add, then the prompt's OK) and ALWAYS cancels; like getRequestForm it refuses any page carrying a bpfButtonBar. tools/reports.ts registers workday_run_report and workday_get_dashboard.
  • src/parse-feedback.ts — the Give Feedback / Get Feedback forms: commit ids, the feedback field, badge options, recipients and template prompts, the question list, and the hidden visibility defaults (which the preview states); plus changeSummary readers for added question rows.
  • src/rich-text.ts — Workday rich text through @chrischall/mcp-utils/rich. Prose a person wrote reads as Markdown (readRichText), labels and prompts as text (readPlainText, which stripHtml now is). Writes go Markdown → toWorkdayHtml, kept to the tags Workday was seen to store (p br ul li b span); what it can't hold is said in words (numbers, link addresses). Only Workday's own quirks live here (&#xa; as a newline, that tag set).
  • src/transport-fetchproxy.ts — wraps @chrischall/mcp-utils/fetchproxy; splitHost maps wd5.myworkday.com → domain myworkday.com + subdomain wd5.
  • src/tools/*.ts — registerXxxTools(server, client) → server.registerTool. review.ts is the review-season surface (team, reviews, feedback, request forms).
  • src/index.ts — wires it with runMcp; brings the bridge up before stdio.

Read the full file on GitHub · 126 lines

Changes

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.

  1. 5d ago Changed · +6 lines · +130 tokens per session 0d7b4bfb765e
  2. 6d ago Changed · +20 lines · +364 tokens per session 2c388c4f89c2
  3. 10d ago Changed · +14 lines · +253 tokens per session 2d0820cdce07
  4. 17d ago Changed · +17 lines · +313 tokens per session 1df80251821e
  5. 1mo ago Changed · +4 lines · -104 tokens per session 608c10b95504
  6. 1mo ago Changed · +6 lines · +123 tokens per session 250ee125fda9
  7. 2mo ago First seen · 59 lines · 1,077 tokens per session scan A 1d3cc53eca10

Subscribe to this mod's changes

workday-mcp CLAUDE.md is an instructions file published in the GitHub repository chrischall/workday-mcp (1 stars, last pushed 3d ago), licensed MIT. It adds 2,156 tokens to every session, about $0.0086 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 2026-08-31.

Related

workday-mcp CLAUDE.md compared with instructions from other repositories

workday-mcp CLAUDE.md next to 6 instructions from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
workday-mcp CLAUDE.mdchrischall/workday-mcpInstructions file13d agoMITA2,156Claude Code
vscode buildNext.instructions.mdmicrosoft/vscodeInstructions file196ktodayMITA9,719GitHub Copilot
next.js AGENTS.mdvercel/next.jsInstructions file144ktodayMITA8,130Codex, OpenCode
vscode oss-third-party-notices.instructions.mdmicrosoft/vscodeInstructions file196ktodayMITA5,685GitHub Copilot
langchain AGENTS.mdlangchain-ai/langchainInstructions file150ktodayMITA4,680Codex, OpenCode
vscode copilot-instructions.mdmicrosoft/vscodeInstructions file196ktodayMITA3,153GitHub Copilot
generative-ai-for-beginners AGENTS.mdmicrosoft/generative-ai-for-beginnersInstructions file124ktodayMITA2,981Codex, OpenCode