Run any Skill in Manus with one click

debug

Stars1

Forks0

UpdatedJune 10, 2026 at 09:48

Use on any bug, test failure, or unexpected behaviour BEFORE proposing a fix — runs the generic evidence-first root-cause loop.

Installation

Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.

Run Skill in Manus

Source

rhanka

rhanka/sentropic

View GitHub Repository View Creator Repositories

Download

Run Skill in Manus

Related occupationsSOC

Based on SOC occupation classification

Software DevelopersComputer and Mathematical Occupations·SOC 15-1252

SKILL.md

readonly

name	debug
description	Use on any bug, test failure, or unexpected behaviour BEFORE proposing a fix — runs the generic evidence-first root-cause loop.

harness/debug

Native sentropic debugging. Open it with harness debug "<symptom>" (records a WorkEvent), then run the loop. No fix is proposed before the root cause is proven by evidence.

The loop

Reproduce — a deterministic, minimal repro. If you can't reproduce it, you can't fix it.
Evidence — gather facts with the tools, never assume: make logs-<service>, make db-query QUERY="…", container state (make ps), grep/diff. Read the actual error, the actual state.
One hypothesis — the single most-likely cause that the evidence supports. Write it down.
One fix — the smallest change that addresses THAT cause. No shotgun edits, no "while I'm here".
Re-observe — rerun the repro. Confirm the symptom is gone AND the evidence changed as predicted.
Regression test — promote the repro into a test (harness/test) so it can't silently return.

Rules

No "probably" / "maybe" / "should be" — verify with a tool before asserting. A claim without evidence is a guess.
Change ONE thing at a time; if the symptom persists, revert it before trying the next.
A failing assertion is data, not noise — read the diff, don't loosen the assertion.
Live surfaces (UI/stream/API/extension) have a dev-lane bootstrap (Playwright/probe targets); use it to observe the real behaviour rather than reasoning in the abstract.

Done

Root cause named + evidenced, minimal fix in place, repro green as a regression test, no collateral change. Hand verification to harness verify.

More from this repository

same repository

adopt

rhanka/sentropic

Use when bringing the harness method to a NEW or non-sentropic repo — maps the repo's conventions into a harness profile so the same scope/branch/verify kernel applies anywhere.

2026-06-101

brainstorm

rhanka/sentropic

Use before any creative/design work (a feature, a package, an evolution) — explores intent and design via the spec-ladder with multi-peer adversarial review BEFORE implementation.

2026-06-101

plan

rhanka/sentropic

Use when you have a spec/decisions and need an executable plan — writes a lot-based BRANCH.md from the template and completes with track for realization state.

2026-06-101

review

rhanka/sentropic

Use when completing a feature, a design step, or before merging — runs a ≥2-peer consensus review and reconciles the findings, instead of a single rubber-stamp pass.

2026-06-101

test

rhanka/sentropic

Use when adding or evolving tests for a feature/bugfix — applies the sentropic test-pyramid, scoped-test loop, environment isolation, and AI-flaky policy.

2026-06-101

using-harness

rhanka/sentropic

Load FIRST in any sentropic repo — establishes the harness method (branch/scope/test/debug/review/brainstorm/plan) and supersedes superpowers for those acts.

2026-06-101

name	debug
description	Use on any bug, test failure, or unexpected behaviour BEFORE proposing a fix — runs the generic evidence-first root-cause loop.

harness/debug

Native sentropic debugging. Open it with harness debug "<symptom>" (records a WorkEvent), then run the loop. No fix is proposed before the root cause is proven by evidence.

The loop

Reproduce — a deterministic, minimal repro. If you can't reproduce it, you can't fix it.
Evidence — gather facts with the tools, never assume: make logs-<service>, make db-query QUERY="…", container state (make ps), grep/diff. Read the actual error, the actual state.
One hypothesis — the single most-likely cause that the evidence supports. Write it down.
One fix — the smallest change that addresses THAT cause. No shotgun edits, no "while I'm here".
Re-observe — rerun the repro. Confirm the symptom is gone AND the evidence changed as predicted.
Regression test — promote the repro into a test (harness/test) so it can't silently return.

Rules

No "probably" / "maybe" / "should be" — verify with a tool before asserting. A claim without evidence is a guess.
Change ONE thing at a time; if the symptom persists, revert it before trying the next.
A failing assertion is data, not noise — read the diff, don't loosen the assertion.
Live surfaces (UI/stream/API/extension) have a dev-lane bootstrap (Playwright/probe targets); use it to observe the real behaviour rather than reasoning in the abstract.

Done

Root cause named + evidenced, minimal fix in place, repro green as a regression test, no collateral change. Hand verification to harness verify.