Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/mishahanin/heading-osnpx agentmods add commands/mishahanin/heading-os/compact-atWrote 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.
[](https://agentmods.dev/commands/mishahanin/heading-os/compact-at)<a href="https://agentmods.dev/commands/mishahanin/heading-os/compact-at"><img src="https://agentmods.dev/badge/commands/mishahanin/heading-os/compact-at.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00017 | $0.00333 |
| Opus 5 | $0.00009 | $0.00167 |
| Sonnet 5 | $0.00003 | $0.00067 |
| Haiku 4.5 | $0.00002 | $0.00033 |
Grade A, and why
compact-at 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 7d 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.
Copies of this mod
1 near-identical copy found in the catalogue:
- compact-at — 92% identical, 6 lines differ
What it actually says
Set the compaction threshold for THIS session. Do nothing else.
The argument is $ARGUMENTS. Read the first token only.
- If the token is a whole number,
status, oroff, run that command:
python scripts/checkpoint-paths.py --compact-at <token>
- If the argument is empty, run the status command:
python scripts/checkpoint-paths.py --compact-at status
- If the token is anything else, print one line:
usage: /compact-at N|status|off. Run no command.
Report the command output in one line. Then stop.
The command refuses a number outside 15-90, and refuses one at or below the session's last rendered fill.
An accepted number also turns unattended on, which turns auto on with it, so
one command is enough and the hook compacts at the threshold instead of asking.
A refusal raises nothing, and status and off raise nothing. When the mode is
already on the command leaves the running stretch alone. Only the operator lowers
it, with /unattended off.
Do NOT write a checkpoint file. Do NOT run /compact. Do NOT continue any task.
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.
- 7d ago First seen · 38 lines · 17 tokens per session scan A d3be98ceb591
compact-at is a command published in the GitHub repository mishahanin/heading-os (11 stars, last pushed yesterday), licensed Apache-2.0. It adds 17 tokens to every session and 333 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 2026-08-30.
Other commands, from other repositories
analyze
Analyze Claude Code transcripts to surface subagent behavior and improvement opportunities.
explain
Generate an interactive one-module explainer for a topic — Riko gathers code scope, Senku plans a 3-5 screen teaching arc, Speedwagon authors the HTML, assembler produces explain-out/index.html.
update-counts
Update all hardcoded counts (skills, agents, marketplace repos/skills) across documentation from filesystem.
color-palette
Generate an accessible colour palette from a hex value.
vdb-scan.template
This prompt was authored for Claude-style slash workflows. In Codex runtime, adapt tool calls as follows.
verify
Run lint, type-check, tests, and a security scan before committing.