DashClaw copilot-instructions.md

A set of GitHub Copilot instructions for DashClaw, a system that controls and records what AI agents do in production.

In plain words
What is it for?
Use it when changing DashClaw’s interface, marketing text, architecture, API-related code, or documentation for agent developers and compliance teams.
Why use it?
It gives Copilot the project’s architecture, audience, design direction, and source-of-truth documents so its code and copy match the product.

Instructions file for GitHub Copilot

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/ucsandman/dashclaw/copilot-instructions
Clone the repo
git clone --depth 1 https://github.com/ucsandman/DashClaw

Made for: GitHub Copilot.

Per session 1,691 This file is loaded in full into every session.
When invoked 1,691 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 $0.01691 $0.01691
Opus 5 $0.00846 $0.00846
Sonnet 5 $0.00338 $0.00338
Haiku 4.5 $0.00169 $0.00169

Measured 2d ago against content hash feae1211dca6, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

DashClaw copilot-instructions.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.

.github/copilot-instructions.md · 66 lines

How it starts

The opening of the file, as written. The whole thing — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.

DashClaw — GitHub Copilot Instructions

DashClaw is a production Next.js 16 governance runtime for AI agents. Copilot should treat the repo's own CLAUDE.md, PROJECT_DETAILS.md, and docs/architecture/runtime-api.md as the source of truth for architecture and API shape. The section below is the canonical design tone — apply it to any UI, marketing, or copy change.

Design Context

Users

DashClaw has two overlapping audiences who see the same surfaces but read them differently:

  • Primary: AI-agent developers and platform engineers. They're integrating DashClaw into Claude Code, Claude Managed Agents, LangChain, CrewAI, OpenAI Agents SDK, custom runtimes, or MCP hosts. They live in terminals, read code more naturally than prose, and judge a product in the first 60 seconds by the quality of its README, SDK ergonomics, and error messages. Their job-to-be-done: "Let my agent act in production without it doing something expensive, irreversible, or embarrassing."
  • Secondary: governance / compliance / security stakeholders. They rarely write code but need to audit agent behavior, approve risky actions, and produce evidence. Their job-to-be-done: "Show me, verifiably, what the agents did, why it was allowed, and who approved anything sensitive."

The context of use is almost always professional, focused, and consequential. Operators open Approvals because an agent is running in prod. Developers open /connect because they're wiring up a live integration. No one is idly browsing. Every pixel should respect that — no decorative filler, no tutorials-for-their-own-sake, no "welcome to your new dashboard" fluff.

Brand Personality

Three words: Serious · Precise · Trustworthy.

DashClaw is a governance runtime. It sits on the critical path between an AI agent's intent and the real world. The visual and verbal tone must match the weight of that position:

  • Voice: direct, technical, declarative. Short sentences. Verbs like intercept, enforce, record, verify. No hype, no exclamation marks, no "unleash your agents."
  • Tone shift: slightly warmer on marketing pages (landing, /connect, docs intros), strictly neutral on operational surfaces (Approvals, Decisions, Policies).
  • Emotional target for operational surfaces: quiet confidence — "things are under control." When an operator opens Approvals, the room should feel like a calm instrument panel, not an alarm board. Status should be obvious at a glance. Brand orange appears only when attention is actually required, not as decoration.
  • Emotional target for marketing surfaces: confident competence. We are the adults in the AI-safety room.

Read the full file on GitHub · 66 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 · 66 lines · 1,691 tokens per session scan A feae1211dca6

Subscribe to this mod's changes

DashClaw copilot-instructions.md is an instructions file published in the GitHub repository ucsandman/DashClaw (296 stars, last pushed 4d ago), licensed MIT. It adds 1,691 tokens to every session, about $0.0085 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 2026-08-30.