push

A decision-making guide for exploring a bold next step after a safe, tested version already exists. It compares the conservative option with a riskier adjacent idea and proposes a small experiment to tell them apart.

In plain words
What is it for?
Use it to evaluate product, business, or project experiments, choose between a safe path and a boundary-pushing path, and define the cheapest test before committing.
Why use it?
It helps avoid both premature risk and endless caution by making the possible upside and failure case explicit. It also identifies the first move that would be difficult to undo.

Skill for Claude CodeCodex

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 skills/megaprompting/torque-loop/push
Any agent
npx skills add Megaprompting/torque-loop --skill push
Clone the repo
git clone --depth 1 https://github.com/Megaprompting/torque-loop

Made for: Claude Code, Codex.

Per session 78 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 586 The whole file, excluding the scripts and references it only reads on demand.
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 $0.00078 $0.00586
Opus 5 $0.00039 $0.00293
Sonnet 5 $0.00016 $0.00117
Haiku 4.5 $0.00008 $0.00059

Measured yesterday against content hash 624af4442d3c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

push 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.

skills/push/SKILL.md · 64 lines

How it starts

The opening of the file, as written. The whole thing — 64 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/ratchet:push — the boundary push

Once the safe version exists, the question changes from "does it work?" to "how much more is on the table?" This command finds the edge where information gain is highest — the bold adjacent move that could multiply the result — without tipping into fantasy.

Step 0 — Load state

ratchet status

Only run this after a safe, tested artifact exists. Pushing an unbuilt idea is not aggression; it is avoidance.

Procedure

Given the current artifact or plan, return:

  1. Safe default path — the conservative version you already have or would ship.
  2. Boundary-pushing path — the boldest adjacent move that could multiply the result.
  3. Why it might work — the mechanism that would make the upside real.
  4. Why it might be stupid — the honest failure case; the way it turns into fantasy.
  5. Cheapest distinguishing experiment — the smallest test that tells you which path is real before you commit to either.
  6. The version to try — your pick, stated.
  7. First irreversible move — the action that commits you and can't be trivially undone.

Rules

  • Not reckless, not timid. Reckless ignores the failure case; timid never leaves the safe path. Aim for the edge where information gain per unit of risk is highest.
  • The boundary path must have a real mechanism, not just ambition. If you can't say why it would work, it's fantasy — mark it so.

Output contract

SAFE DEFAULT: <path>
BOUNDARY PATH: <bold adjacent move>
WHY IT WORKS: <mechanism>
WHY IT'S STUPID: <honest failure case>
DISTINGUISHING EXPERIMENT: <cheapest test>
VERSION TO TRY: <pick>
FIRST IRREVERSIBLE MOVE: <the committing action>

Serialize

Record the push as a decision with a built-in reversal tripwire:

ratchet state append decisions '{"choice":"push: <boundary path>","rejected":"safe default","tripwire":"<signal the experiment fails>"}'
ratchet state set phase decide

Next: /ratchet:build the distinguishing experiment first. The experiment is cheaper than the commitment — run it before the irreversible move.

Read the full file on GitHub · 64 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. yesterday First seen · 64 lines · 78 tokens per session scan A 624af4442d3c

Subscribe to this mod's changes

push is a skill published in the GitHub repository Megaprompting/torque-loop (5 stars, last pushed 1mo ago), licensed MIT. It adds 78 tokens to every session and 586 once invoked, about $0.0004 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.

Related

Other skills, from other repositories

new-plugin

Factory line for adding a new HAR verification plugin (like playwright or rocketsim) for any framework — research the framework docs, build the template under src/templates/plugins/, register it everywhere, validate on a real repository, and open a PR. Use when asked to add/create a plugin, plugin template, or…

os-factory/har · 89 tokens

factory-line

Factory line for executing one station of a declared multi-station program — read the installed line bundle (har line status), plan parallel work into isolated HAR slots, run the cumulative gate with har line gate, and hand off for human review. Use when asked to "run a factory line", "run the next station", "execute…

os-factory/har · 101 tokens

v1-milestone

Factory line for executing one milestone of the HAR v1.0.0 refactor (epic os-factory/har#225) — plan the wave of parallel subagents, implement each issue in its own HAR slot, ship stacked PRs, run the fixture-e2e milestone gate, and hand off for review. Use when asked to "run the next v1 milestone", "work on v1.0.0"…

os-factory/har · 110 tokens

ctx

Codebase intelligence and evidence-driven governance with the indexed ctx CLI. Use when exploring an unfamiliar repository, locating symbols or callers, checking for existing implementations, estimating change impact, enforcing architecture rules, scoring a branch, finding hotspots or duplication, or analyzing…

agentis-tools/ctx · 60 tokens

ctx

Codebase intelligence and evidence-driven governance with the indexed ctx CLI. Use when exploring an unfamiliar repository, locating symbols or callers, checking for existing implementations, estimating change impact, enforcing architecture rules, scoring a branch, finding hotspots or duplication, or analyzing…

agentis-tools/ctx · 60 tokens

golden-rss

Use when testing the rss golden build.

yusufkaraaslan/Skill_Seekers · 12 tokens