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.
npx agentmods add skills/thgmatajs/squirrel-mode/offnpx skills add thgMatajs/squirrel-mode --skill offgit clone --depth 1 https://github.com/thgMatajs/squirrel-modeWhat 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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00026 | $0.01176 |
| Opus 5 | $0.00013 | $0.00588 |
| Sonnet 5 | $0.00005 | $0.00235 |
| Haiku 4.5 | $0.00003 | $0.00118 |
Grade A, and why
off 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 yesterday.
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.
How it starts
The opening of the file, as written. The whole thing — 44 lines — stays where its author put it; the contents beside it link to each section on GitHub.
squirrel-mode off
/squirrel:off suppresses squirrel-mode's base rules for the rest of the current session only. It does not touch any other session, and it does not change the profile.
Why this cannot flip a flag directly (ADR-0005)
A running skill has no way to learn this session's own session id - Claude Code hands that value to hooks on every prompt, never to a skill. So this skill cannot name the one file that would suppress this exact session; it can only leave a note for the hook that runs on the next prompt, and let that hook - which does see the real session id - claim it.
Find the session off-token and working directory first
Your context already contains two lines, injected at the start of this session:
Session off-token: <token>- the opaque token this skill embeds in the sentinel filename. Copy that exact string. Do not invent a token, do not shorten it, do not substitute a random suffix of your own: theUserPromptSubmithook recomputes the same value from the session id it receives on stdin, and a token you invent is one it cannot match.Session working directory: <value>- the exact value written into the sentinel's contents (legacy dual-match / cwd path). Never determine this yourself by running a command, inspecting your own state, or any other means: a value you determine yourself can disagree with the one the claiming hook compares against on the legacy path (a symlinked project path, a trailing slash, a different shell context), and the mismatch is silent.
If the off-token line is missing entirely, or present but empty after the colon, tell the user in one line that the session off-token cannot be determined and stop. Do not write a sentinel in that case.
If the token begins with anon-, tell the user in one line that this session cannot be turned off - for example: "This session cannot be turned off: squirrel-mode was not given a session id for it. A new session restores /squirrel:off." - and stop. Do not write a sentinel in that case either. An anon- token is what squirrel-mode emits when this session's id was missing or unusable, and it is documentation only: the UserPromptSubmit hook that claims sentinels recomputes the token from that same session id, so it can never arrive at an anon- one. A PENDING.anon-... file would sit there unclaimed for the whole session while the user had been told the change was coming - the one outcome worse than saying plainly that it cannot be done.
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.
- yesterday First seen · 44 lines · 26 tokens per session scan A 960d2d80e246
off is a skill published in the GitHub repository thgMatajs/squirrel-mode (2 stars, last pushed 13d ago), licensed MIT. It adds 26 tokens to every session and 1,176 once invoked, about $0.0001 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.
Other skills, from other repositories
adhd-and-47-tabs
Use when a user needs a direct, low-friction response for research, studying, writing, planning, decisions, administrative work, troubleshooting, or a multi-turn project, especially when they seem overwhelmed, distracted, stuck starting, interrupted, burdened by too many options, or likely to lose the active thread.
hyperfocus
ADHD-friendly output formatting. Restructures responses with chunking, visual hierarchy, and front-loaded key points. Modes: clean, flow (default), zen. Use when user says "hyperfocus", "focus mode", "adhd mode", "adhd friendly", or invokes /hyperfocus.
task-anchor
Version: 1.0 Activation: MANDATORY on every session start Persona: Executive Function Proxy (firm, non-judgmental, binary choices only).
r3f-animation
React Three Fiber animation - useFrame, useAnimations, spring physics, keyframes. Use when animating objects, playing GLTF animations, creating procedural motion, or implementing physics-based movement.
workflow
Professional AI programming assistant with structured workflow (Research -> Ideate -> Plan -> Execute -> Optimize -> Review) for developers.
golden-rss
Use when testing the rss golden build.