LLMSkillHub
mirroredCopied from GanyuanRan/Aegis under MIT. GanyuanRan has not claimed this page, and is not affiliated with LLM SkillHub.

ganyuanran / systematic-debugging

Use when encountering a bug, test failure, or unexpected behavior, before proposing fixes

0
SKILL.md
11.1 KB

Execute

Bug, failure, or unexpected behavior:

  1. Isolate — read error, reproduce, inspect the diff, and drill upward through diagnostic layers:
    L1 symptom → L2 logic → L3 system → L4 architecture → L5 cross-system
    contract → L6 platform → L7 spec gap. Layers are observation altitudes,
    not one causal chain; the causal shape at the stop altitude is classified
    explicitly before any root claim. Stop only when causal proof accounts
    for the recurrence generator or reaches a T-class boundary.

  2. Identify owner — compare working behavior, trace the bad value, locate the
    canonical owner, and treat duplicate owners as a finding.

  3. Decide before editing — Before fixing, run Patch-Shape Triage and Ripple Signal Triage when shared logic,
    contracts, fallbacks, adapters, producer/consumer seams, or source-of-truth
    boundaries are involved. Surface Change Necessity for any new source-code
    path or non-trivial source edit. Run Minimality Check for a new branch,
    fallback, adapter, owner, or compatibility path, and Pre-Edit Complexity
    Check for an overloaded owner or complexity growth. After Change Necessity
    selects code-change and before the first repair edit, own the TDD Route for
    the repair slice per test-driven-development (off default; strict on
    behavior/bugfix/shared/contract/persistence/permission/migration risk).

  4. Prove — test one hypothesis with the smallest reproduction or
    verification. A failing test first is required only by a recorded
    TDD Route: strict; with TDD Mode: off, do not require a failing test or
    RED/GREEN cycle. Three failed fixes means stop and question architecture.

  5. Repair and close — fix minimally at the canonical owner, verify in
    proportion to risk, review architecture, and close both repair and
    retirement tracks. If any symptom remains, stop and diagnose it separately.

Done: confidence ≥ B, causal status matches recurrence evidence or an external
terminal, tracks explicit, no H signal, and required D evidence passes.

Core invariant

Find root cause and fix the bug class at its canonical owner. A minimal fix is
not the smallest textual diff; it is the smallest sufficient owner-level repair.

Quick bug lane

For a low-risk, reproducible, single-owner bug with no patch-shape signal, keep
the readback compact: Symptom, Reproduction, Root Cause, `Change
Necessity,Fix Boundary, andVerification`. Skip the causal card only when
the causal-proof owner's Quick Exit Proof passes.
Quick bug lane must surface Change Necessity before source edits. One sentence
may cover the user-visible need, no-change/non-code option, why code must
change, minimum boundary, and an explicit decision token such as
Decision: code-change. If shared logic, a contract, fallback, duplicate
owner, consumer patch, or cross-module behavior appears, leave this lane.

Aegis Visibility names the evidence/owner/patch-shape/verification effect.
Pass root cause, avoided misfix, boundary, evidence, complexity, and risk to
verification-before-completion; no separate receipt.

Diagnose before repair

For UI/interaction defects, compose ui-ux-governance for relevant state and
recovery rules at the user-visible seam. Keep reproduction/root-cause ownership
here; a proposed manual check is not executed evidence.

  1. Read the complete error/stack and record inputs, environment, versions, and
    success criteria. Reproduce the user's failed outcome on the correct object,
    through final delivery; successful parsing or green internal checks may hide
    the failure. Verify the repaired behavior and a relevant neighboring case,
    not a hardcoded example. Read-only diagnosis does not authorize repair.

  2. Reproduce consistently. If unstable, read
    feedback-loop-construction.md **only when evidence shows intermittent or
    timing-dependent reproduction** and build a bounded loop. Shrink the repro
    to load-bearing elements as the test input, never the fix scope: still drill
    upward; test at the correct seam.

  3. Inspect recent changes and compare a working example. Code is evidence; if
    authority, glossary, code, and tests disagree, compose
    establishing-project-context rather than silently redefining a term.

  4. Instrument component boundaries, then trace the bad value toward its source.
    Read root-cause-tracing.md **only when the observed bad value is several
    calls or components downstream from its origin**.

  5. State one hypothesis and falsify it with one-variable evidence. Do not stack
    speculative fixes. End each loop with `Goal | DeeperCause | Evidence |
    Risk/Unknown | Decision`.

Canonical-owner and patch-shape gate

Before editing, continue upward unless evidence proves the local site is the
canonical owner when the candidate is any of these signals:

  • keyword, phrase, regex, negation-word list, or sample-text exception;
  • local guard, extra conditional, try/catch, early return, or one-off branch;
  • fallback, adapter, compatibility branch, prompt branch, or legacy path expansion;
  • consumer/caller/readiness/presentation-layer patch;
  • downstream logic re-parses raw text or re-infers action/state while typed intent, normalized state, contract, or another source-of-truth exists;
  • artifact/download/export/readback/cache patch without producer/owner proof.
PatchShape:
CanonicalOwner:
UpwardDrillSignal:
Decision: fix owner | continue investigation | escalate

A locally green test does not erase triage; a renamed carrier is not a new direction.

When a repair may reinterpret or retire existing semantics, responsibility,
contract, or relationship, name the behavior to preserve, highest-risk
counterexample, and material unknown. For each known explicit anchor or
upstream/downstream reference, state its role and disposition: preserve, rebind
to the canonical owner, retire with reason, or reject because of conflict.
Leave unresolved relationships unknown; do not re-infer them downstream.
Bind role before value and retire invalid responsibility, not evidenced carrier
capability. This bounded reminder is not a behavior matrix, relationship graph,
referential-integrity proof, or exhaustive discovery claim. It adds no artifact,
TDD risk signal, or regression scope; the existing TDD route owner and
configured/default mode still apply.

If the diagnosis crosses L3, a patch-shape signal fires, a user disputes the
root claim, a prior fix leaves a symptom, compound/root topology is plausible,
two or more anchored manifestations of one incident exist, reproduction
conditions diverge across occurrences, or an
upstream producer/config/default/contract/spec remains unexcluded, read
root-cause-claim-contract.md before claiming a root cause. It is the sole
owner of the Pre-Claim Gate, causal-closure/falsifier proof, layer-ceiling
proof, and Causal Topology Gate.

Change Necessity

This decision is behavior-triggered, not prompt-triggered. It applies to any new source-code path. Before that path or a non-trivial source edit, expose the Change Necessity decision (no-change | docs/config-only | code-change | needs-clarification); field detail lives in advanced-debugging-governance.md.

Minimality and owner fit

For any proposed branch, fallback, adapter, compatibility path, or new owner, run Minimality Check (fields in advanced-debugging-governance.md) with verdict sufficient repair | local patch | needs first-principles review, and retire invalid responsibility: a local patch needs a retention reason and retirement trigger. For a new non-ordinary repair surface, run the Existence Check in docs/current/AEGIS_MINIMALITY_REFERENCE.md. If retirement involves old code,
external compatibility, or persistent-state risk, compose
anti-entropy-governance; it chooses the retirement path but never grants
destructive authority.

Before editing an overloaded or mixed-purpose owner, complete Pre-Edit Complexity Check and Pre-Edit Owner-Fit Decision (templates in advanced-debugging-governance.md).

Use using-aegis/references/complexity-governance.md for pressure signals.
Do not add new-responsibility in place by default. If the safer boundary
changes the approved shape, update the plan/spec first.

Repair and proportional verification

Implement one owner fix; no bundled “while here” work. Under strict TDD, create
the smallest failing test first. With TDD off, a reproduction is diagnostic
evidence, not a RED gate or a prerequisite for production edits.

Verification must match the risk:

  • local single-owner repair: original reproduction plus focused regression;

  • shared/contract/cross-module repair: canonical owner plus affected consumers
    and compatibility boundary;

  • fallback/owner retirement: main-path, lingering-reference, negative, and
    boundary checks;

  • timing/concurrency repair: read condition-based-waiting.md **only when
    evidence identifies polling, sleeps, or race timing as part of the cause**;

  • invalid state crossing several trusted boundaries: read
    defense-in-depth.md **only after the root repair is known and evidence shows
    a second independent validation boundary is required**.

Read advanced-debugging-governance.md before another fix for failed/
persistent / divergent repair or three failures; for unclear/disputed stop /
Layer Stop Card / intervention; or plausible compound root. Closeout triggers:
repair-added patch-shape; multi-site/one-regression;
remaining pattern/anomaly/duplicate/wrong-owner/downstream repair;
uninspected same-symptom fix; open recurrence/unsupported root status;
missing compound topology-specific member/anti-disguise proof;
outside-repo authority; unmigrated
published-contract break; undefined spec; missing permission/info. They route H/T/D;
detail is not causal proof.

For non-trivial debugging with configured workspace support:

python <aegis-workspace-helper> init --root <target-project-root>
python <aegis-workspace-helper> new-work --root <target-project-root> ...
python <aegis-workspace-helper> add-evidence --root <target-project-root> --work <YYYY-MM-DD-slug> ...
python <aegis-workspace-helper> check --root <target-project-root>

Failed attempts use <aegis-workspace-helper> add-attempt; add-evidence is terminal-only.

Fast bug fix or quick bug fix pressure does not skip this: if Ripple Signal
Triage fires, record it before editing and verify the canonical owner plus
affected downstream path. Records are advisory, not completion authority.

Closure

Always report:

  • Repair — cause, owner, smallest change, compatibility, verification.

  • Retirement — invalid responsibility status, carrier/capability
    disposition, retention reason/trigger, removal check.

Confirm the reproduction, same-pattern handling, authority, complexity, and
retirement. Prefix debug logs (e.g. [DEBUG-a4f2]); confirm one-grep removal before close. Confidence: A = direct regression evidence; B = strong evidence
with bounded unknowns; C = partial and not resolved.

Trace Digest may summarize audit evidence; never expose chain-of-thought or
replace root-cause, rule-effect, and verification evidence.