dearuser CLAUDE.md

A Claude Code instruction file containing project-wide rules and notes for the Dear User application.

In plain words
What is it for?
Guiding work on Dear User’s collaboration diagnosis, feedback features, local database, slash commands, MCP tools, and process startup.
Why use it?
It keeps the agent aligned with the product’s scope, local-only data handling, command-to-tool mappings, and known process fixes.

Instructions file

Install

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.

agentmods
npx agentmods add instructions/bleedmode/dearuser/claude-md
Clone the repo
git clone --depth 1 https://github.com/bleedmode/dearuser
Per session 1,459 This file is loaded in full into every session.
When invoked 1,459 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
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 $0.01459 $0.01459
Opus 5 $0.00730 $0.00730
Sonnet 5 $0.00292 $0.00292
Haiku 4.5 $0.00146 $0.00146

Measured 2d ago against content hash 0d5c934704b3, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

dearuser 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 2d 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 · 84 lines

How it starts

The opening of the file, as written. The whole thing — 84 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Dear User — project-wide notes

Globale regler i ~/.claude/CLAUDE.md. MCP-internt i mcp/CLAUDE.md. Her ligger learnings der gælder Dear User-produktet på tværs.

Produkt-scope

Menneske↔agent samarbejdsanalyse. Kører lokalt, ingen data forlader maskinen. Live: dearuser.ai. Lokal DB: ~/.dearuser/dearuser.db (SQLite, 6 tabeller: du_migrations, du_agent_runs, du_recommendations, du_score_history, du_findings, du_finding_events — auto-oprettes).

Scope: Diagnose + feedback loop. IKKE task management, research eller dashboard — det er PVS OS.

Slash-command ↔ MCP-tool mapping

  • /dearuser-collabmcp__dearuser__collab
  • /dearuser-healthmcp__dearuser__health
  • /dearuser-securitymcp__dearuser__security
  • /dearuser-wrappedmcp__dearuser__wrapped
  • /dearuser-onboardmcp__dearuser__onboard
  • /dearuser-historymcp__dearuser__history
  • /dearuser-helpmcp__dearuser__help
  • /dearuser-feedbackmcp__dearuser__feedback

Wrapped can be shared publicly via mcp__dearuser__share_report — handled from within the wrapped flow, not a separate slash-command. Collab/health/security reports are local-only.

Dashboard: http://localhost:7700 — Config: ~/.dearuser/config.json

Lessons fra Dear User-udvikling

MCP tool descriptions har 2000-char hard cap (test-enforced)

mcp/src/tool-descriptions.test.ts enforcer length <= 2000 per tool. Flere af descriptions ligger få chars under grænsen — health-tool var 1998/2000 da vi ramte den 2026-04-27.

Why: Anthropic-registry og MCP-clients viser fuld description; lang copy bliver upraktisk for både agent-discovery og humans. Test fanger drift.

How to apply: før du tilføjer copy til en eksisterende tool description, mål nuværende length først:

node -e "const m = require('fs').readFileSync('src/index.ts','utf8').match(/server\.tool\(\s*'<name>',\s*\`([^\`]+)\`/); console.log(m[1].length)"

Hvis margin <50 chars: trim eksisterende inden du adderer. Foretrukne trim-targets er Examples: (lavt info-tab) og redundans i BehaviorWhat this tool does NOT do.

Read the full file on GitHub · 84 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. 2d ago First seen · 84 lines · 1,459 tokens per session scan A 0d5c934704b3

Subscribe to this mod's changes

dearuser CLAUDE.md is an instructions file published in the GitHub repository bleedmode/dearuser (0 stars, last pushed 26d ago), licensed MIT. It adds 1,459 tokens to every session, about $0.0073 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.