agent-severance

agent-severance is a cursor rule for Cursor from mohitagw15856/pm-claude-skills. It costs 104 tokens per session (1,261 once invoked), scanned A, original, MIT.

An offboarding checklist for an AI agent or bot, similar to removing an employee’s access when they leave. It covers what the agent could access, what it remembered, what it did, and what a replacement needs to know.

In plain words
What is it for?
Retiring an agent, changing vendors, ending an AI trial, or replacing a bot. It helps inventory access and memory, revoke credentials, record what was kept or deleted, and document responsibilities.
Why use it?
Simply deleting a chat or turning off a bot may leave credentials, scheduled tasks, integrations, or stored information active. It helps close those access and handover gaps.

Cursor rule for Cursor

Written for Cursor: a Cursor rule (.mdc).

Good fit Retiring an agent, changing vendors, ending an AI trial, or replacing a bot. It helps inventory access and memory, revoke credentials, record what was kept or deleted, and document responsibilities.

Compare 6 cursor rules from other repositories ↓
Install with agentmods
npx agentmods add rules/mohitagw15856/pm-claude-skills/agent-severance
About the project

PM Skills is a collection of plain-Markdown instructions that teach AI assistants structured methods for handling professional, personal, and life-admin tasks. People use it with Claude, ChatGPT, Gemini, Cursor, Codex, and other supported agents for work such as writing product requirements, reviewing documents, or planning difficult situations.

mohitagw15856/pm-claude-skills · 1,345 stars · on GitHub · mohitagw15856.github.io

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.

Clone the repo
git clone --depth 1 https://github.com/mohitagw15856/pm-claude-skills

Made for: Cursor.

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 agent-severance

README.md
[![agentmods](https://agentmods.dev/badge/rules/mohitagw15856/pm-claude-skills/agent-severance/github.svg)](https://agentmods.dev/rules/mohitagw15856/pm-claude-skills/agent-severance)
Your own site
<a href="https://agentmods.dev/rules/mohitagw15856/pm-claude-skills/agent-severance"><img src="https://agentmods.dev/badge/rules/mohitagw15856/pm-claude-skills/agent-severance/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 agent-severance

Your own site · 80×15
<a href="https://agentmods.dev/rules/mohitagw15856/pm-claude-skills/agent-severance"><img src="https://agentmods.dev/badge/rules/mohitagw15856/pm-claude-skills/agent-severance.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 104 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,261 The whole file, excluding the scripts and references it only reads on demand.
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 $0.00104 $0.01261
Opus 5 $0.00052 $0.00630
Sonnet 5 $0.00021 $0.00252
Haiku 4.5 $0.00010 $0.00126

Measured 10d ago against content hash 87dda0dca083, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

agent-severance 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 10d 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.

exports/cursor/pm-2027/agent-severance/agent-severance.mdc · 114 lines

How it starts

The opening of the file, as written. The whole thing — 114 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Agent Severance Skill

Orgs learned employee offboarding the hard way: the contractor whose VPN worked for a year after the contract, the shared password nobody rotated. Long-lived agents recreate every one of those failure modes with worse logging — an agent accumulates credentials, memory, integrations, scheduled jobs, and undocumented responsibilities, and then one day it's "turned off" by deleting a chat window while its API keys live on. This skill runs the severance properly: know what it had, keep what's valuable, kill what's live, and hand over what it did.

What This Skill Produces

  • An inventory: everything the agent could touch (credentials, tools, data stores, channels), everything it knew (memory, context files, fine-tuning or instructions), and everything it did on a schedule
  • An access-revocation table with owner and verification step per row — revoked isn't revoked until someone confirmed the key is dead
  • A memory disposition record: exported / retained (where, why, how long) / purged (how verified) — the part compliance will ask about in 2027
  • A successor handover: the agent's actual duties, including the undocumented ones users discovered, for whoever inherits them

Required Inputs

Ask for (if not already provided):

  • The agent: platform, what it was for, how long it ran, who owned it
  • Known integrations and credentials (then treat the list as incomplete on principle — the inventory step hunts for the rest)
  • Why it's being offboarded (vendor switch, pilot ended, incident, cost) — incident-driven severance changes the order: revoke first, inventory second
  • What must survive: memory worth exporting, workflows someone still needs

Process

  1. Inventory before touching anything (unless incident — then revoke first). Hunt beyond the known list: API keys and OAuth grants · service accounts · webhook URLs pointing at it · scheduled/cron jobs it ran · channels it posted in · data stores it read or wrote · other agents that called it (the A2A dependencies nobody documented) · what its memory contains, including personal data.
  2. Decide memory disposition per store, not wholesale. Export what has value (decisions log, learned context) to an owned location; name a retention owner and period for anything kept; purge the rest and record how purged (vendor deletion request ≠ deleted — note what the vendor actually promises). Personal data follows your privacy policy's deletion rules, flagged explicitly.
  3. Revoke with verification. Every row gets: the credential, who revokes it, and the test that proves it's dead (the call that now fails). Rotate any shared secrets the agent ever held — its copy dying doesn't kill the copies.
  4. Write the honest handover. What it was supposed to do, what it actually did (ask its users — there are always undocumented duties), open threads mid-flight, and the workflows that will silently break next Tuesday when it stops.
  5. Announce the death. One message to the channels it served: it's gone, here's who/what replaces it, here's where its exported memory lives.

Read the full file on GitHub · 114 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. 10d ago First seen · 114 lines · 104 tokens per session scan A 87dda0dca083

Subscribe to this mod's changes

agent-severance is a cursor rule published in the GitHub repository mohitagw15856/pm-claude-skills (1,345 stars, last pushed 2d ago), licensed MIT. It adds 104 tokens to every session and 1,261 once invoked, about $0.0005 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.