The Desk Authority Map for agent teams (free sample)
by PaperChaseWebb
One template that says what each agent may do without you, what it must ask for, what it never does and what proof it owes, with an example row and an audit prompt.
- Price
- Free
- Difficulty
- Beginner
- Setup
- 5 min
- Format
- .md file
desk-authority-map-for-agent-teams-free-sample.md · 2 KB · direct download, no signup
What you end up with
Each agent's limits are written in one table you can check, and no agent signs off its own work.
What you get
- The Desk Authority Map template
- An example row
- Five rules
- An audit prompt
Who it is for
Anyone running agents who has never written down what each one is allowed to do alone.
Works with
Solid chips: the agent is named in this listing's own tools list. Dashed chip: Plain Markdown, so it also pastes into other agents such as Grok Bot, Hermes, OpenClaw or Muse. We list those as compatible by format only, not as tested.
Tools needed
Cost to run: Free.
Time to result: Five minutes per agent
Sample prompts, free to use
Read this authority map: [paste]. Flag any action under Does alone that cannot be undone, any agent that checks its own work, and any Evidence cell a person could not verify in two minutes. Return a corrected map.
Run it again whenever you add an agent or a connector.
Free preview of the file
# Free sample: the Desk Authority Map From Governed Agent Teams by PaperChaseWebb. One template that answers the question most agent setups skip: what may each agent do without you? An org chart says who reports to whom. An authority map says who may commit you to what. Fill one row per agent and link it from each agent's instruction file (Claude Code subagent, Codex AGENTS.md or Hermes skill). ## Template | Agent | Job in one line | Does alone | Asks first | Never does | Evidence it owes | Checked by | |---|---|---|---|---|---|---| | [name] | [one line] | [undoable actions] | [actions, with the card it shows] | [actions] | [what proves done] | [a different agent or person] | ## Example row | Agent | Job in one line | Does alone | Asks first | Never does | Evidence it owes | Checked by | |---|---|---|---|---|---|---| | Builder | Implements one planned task in a branch | Edit code in its branch, run tests | Add a dependency, change a schema | Push to main, change CI secrets | Diff, pasted test output, a handoff packet | Reviewer | ## Five rules 1. Anything under "Does alone" must be undoable. 2. "Evidence it owes" must be something a person can check in two minutes. 3. No agent checks its own work. 4. Every "Asks first" action says what, why and what happens on a no. 5. Review the map whenever you add an agent or a connector. ## Audit prompt ```text Read this authority map: [paste]. Flag any action under Does alone that cannot be undone, any agent that checks its own
Worth knowing
- A written limit only holds if the tool's permission settings match it.
Questions
What is in the full pack?
7 subagent personas for Claude Code with Codex and Hermes versions, three role risk classes, a handoff packet, an evidence standard, a decision ledger, a reach inventory, 12 tests and 14 prompts.
Details
Version 1.0 · Updated Oct 26, 2026 · Markdown version of this page
License: Personal and commercial use. Download terms
Related

Governed Agent Teams: subagent personas, authority maps and handoffs
A Markdown pack for running a team of coding and research agents with clear limits: 7 subagent personas for Claude Code with Codex AGENTS.md and Hermes skill versions, a Desk Authority Map, three role risk classes, a handoff packet the receiver verifies, an evidence standard, a decision ledger and a reach inventory.

Ask-first rules for AI agents: hard nos, approval cards and a test set
Paste-in rules that make an agent ask before it sends, spends, posts or deletes, define what a real yes is, and stop retries after a no, plus 12 test scenarios to check it.

Solo-founder agent ops playbook: run your business with agents without losing control
Decide which jobs go to agents, how much power each gets, one house rules file they all follow, a daily brief, a weekly review, cost caps and a runbook for when an agent gets it wrong.