AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Audit The Premise

skill-ariaxhan-the-agent-library-audit-the-premise · by ariaxhan

>-

No reviews yet
0 installs
17 views
0.0% view→install

Install

$ agentstack add skill-ariaxhan-the-agent-library-audit-the-premise

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-ariaxhan-the-agent-library-audit-the-premise)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
29d ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Audit The Premise? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Audit the Premise

Most tasks arrive with a hidden claim attached: that something is broken, that a fact holds, that the fix is small. Act on the claim and you can spend hours solving a problem that was never real, or confidently break something that was fine. This checks the claim first.

Scale it to stakes: skip it for trivial, well-specified asks; always run it before anything hard to undo.

1. Treat the framing as a hypothesis

Before doing anything, ask the one question that collapses wasted work:

> What single thing, if false, makes this whole task pointless?

Name that load-bearing assumption explicitly. The request is a claim about reality, not reality itself.

2. Check the load-bearing facts against the real state

Look at the actual current state end-to-end before changing anything:

  • "It's broken" → reproduce the break yourself.
  • "It's been down since Tuesday" → check when it actually changed.
  • "Just flip the existing thing" → confirm the thing exists and does what's assumed.

Read-only investigation is cheap. It is what separates a safe two-line change from a confidently-broken big one.

3. If reality contradicts the framing, stop

The contradiction is not an obstacle: it is the finding. When what you see disagrees with what you were told:

  • Do not silently comply and build on the false premise.
  • Do not silently "fix" reality to match the request either.
  • Halt and surface the gap. Never tidy, overwrite, or recreate something on a hunch that it's wrong.

4. Proceed on the truer question and log the disagreement

Once the real state is clear, work the actual problem, which is often smaller and different from the one asked. Write down the disagreement in three lines:

  • What you were asked.
  • What's actually true.
  • What you did instead, and why.

A wrong assumption that never gets recorded comes back and gets believed again months later. Recording it is what stops the loop.

5. Revise the plan

Restate the task as it really is now. Usually it shrinks: a smaller change, a different target, or sometimes "nothing to do here: the premise was false."

Output

Return:

  • The load-bearing assumption you tested.
  • What the real state turned out to be, with how you checked it.
  • A disagreement note (asked / true / did instead) when the premise was wrong.
  • The revised, premise-corrected plan, or a clear statement that no action is needed.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.