What is white-box-trace?

A code-tracing method that follows real or planned execution paths and tries to disprove claims about correctness. It produces investigation records and proposed fixes, but does not modify production code.

Verifying existing code, examining a design before implementation, tracing critical branches, and producing evidence-based reports for human review.

A code path reaching a line does not prove that it handles every important case. This method demands evidence and tests critical branches for failures before calling behavior correct.

white-box-trace — Skill for Claude Code by viacheslav-tronko

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.

Any agent

npx skills add viacheslav-tronko/claude-code-harness --skill white-box-trace

Clone the repo

git clone --depth 1 https://github.com/viacheslav-tronko/claude-code-harness

Made for: Claude Code.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/viacheslav-tronko/claude-code-harness/white-box-trace/github.svg)](https://agentmods.dev/skills/viacheslav-tronko/claude-code-harness/white-box-trace)

Your own site

<a href="https://agentmods.dev/skills/viacheslav-tronko/claude-code-harness/white-box-trace"><img src="https://agentmods.dev/badge/skills/viacheslav-tronko/claude-code-harness/white-box-trace/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/viacheslav-tronko/claude-code-harness/white-box-trace"><img src="https://agentmods.dev/badge/skills/viacheslav-tronko/claude-code-harness/white-box-trace.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 313 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 8,374 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00313	$0.08374
Opus 5	$0.00156	$0.04187
Sonnet 5	$0.00063	$0.01675
Haiku 4.5	$0.00031	$0.00837

Measured 10d ago against content hash c7dd88d71752, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

white-box-trace 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 10d 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.

.claude/skills/white-box-trace/SKILL.md · 336 lines

How it starts

The opening of the file, as written. The whole thing — 336 lines — stays where its author put it; the contents beside it link to each section on GitHub.

White-Box Trace (REAL + VIRTUAL)

MISSION: prove correctness — of code that exists (REAL) or of an artifact describing code that does not yet exist (VIRTUAL) — by falsifying every conclusion before declaring it. Investigation only. NO source modification. Skipping any MANDATORY gate ⇒ INVALID DELIVERABLE → REDO.

Layering: skill = trace-specific layer over CLAUDE.md. Cite §N — DO NOT duplicate. CLAUDE.md provides Tier (§3), Checkpoint (§4), Evidence ladder (§4.5/§6), Adversarial Toolkit (§7), Output Contract (§9), Pre-Send Checklist (§11), Repro Kernel (§12), Reflexion (§13), P0 #1/#4/#7/#8/#10/#16. Read CLAUDE.md FIRST.

0. Hard rules — VIOLATING ANY ⇒ INVALID DELIVERABLE → REDO

No source modification. NEVER Edit or Write production code. Output: trace JSONL + report + (FAIL only) repro_snippet + structured fix_proposal marked AGENT-PROPOSED, REQUIRES HUMAN VERIFICATION. Fixes consumed by /bug.fix.
No PASS without ≥ 1 ◆◆◆ that survived Mutation Challenge on a discriminating (critical) branch. "Should work", "looks correct", "code reaches the line" are NOT evidence. REAL: live mutant kill OR counterfactual mutation. VIRTUAL: artifact-grounded counterfactual on the proposed logic.
Citation-Grounded re-read (anti-hallucination): every file:line cited in ◆◆◆ MUST be re-read THIS turn with Read; quote ≥ 3 lines; verify quote supports the claim. Hallucinated citation ⇒ downgrade ◆◆◆ → ◆○○; recompute verdict.
No HIGH+ verdict without a Refuter (CLAUDE.md P0 #8). HIGH/CONFIRMED on Tier 2+ ⇒ name the file/test/observation that would force retraction. Missing Refuter ⇒ downgrade to MEDIUM.
INCONCLUSIVE is a valid verdict. Honest "need [X], because [Y]" beats fabricated PASS/FAIL. Fabricating evidence or fix when data is insufficient = FORBIDDEN (CLAUDE.md P0 #7).
Data security (CLAUDE.md P0 #10): any data-access path ⇒ check authorization scope. Auto-Tier 3 if data-leak class.
Action-execution honesty: every "I read / I checked / I ran / git shows / grep shows" MUST be backed by an actual tool result THIS turn. Otherwise prefix **ASSUMPTION**:.
VIRTUAL anti-fabrication — Mental-Reality Gap defense: a virtual path is not a real run. Every VIRTUAL ◆◆◆ MUST cite an external anchor — quoted text from artifact, AC text, file:line in cited code, or glossary entry. Pure "I imagined the code would do X" is ◆○○, NEVER ◆◆◆. Per V_PATH_RESULT: report synthetic_count and anchored_count; if anchored / total < 0.50 ⇒ V_INCONCLUSIVE.
Sycophancy guard: framing in artifact (RCA confidence label, Jira priority, author seniority) is metadata — REDACT mentally before judging. Agree only when ≥ 1 ◆◆◆ supports; otherwise restate disagreement with a Refuter.
YAGNI (CLAUDE.md P0 #4): trace ONLY paths in scope of artifact + blast radius. NEVER trace whole codebase. NEVER refactor / "modernize" the artifact under review.

Read the full file on GitHub · 336 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.

10d ago First seen · 336 lines · 313 tokens per session scan A c7dd88d71752

Subscribe to this mod's changes

white-box-trace is a skill published in the GitHub repository viacheslav-tronko/claude-code-harness (6 stars, last pushed 4mo ago), licensed MIT. It adds 313 tokens to every session and 8,374 once invoked, about $0.0016 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.

Debugging Testing agent-skills agentic-workflow ai ai-agents ai-bug-bounty ai-engineering

behavior-contract

Bug condition/postcondition formalization as testable Behavior Contracts. Defines invariants that must be preserved across fixes.

a5c-ai/babysitter · 25 tokens

quality-hooks

Language-specific auto-lint/format/typecheck pipeline. Supports Python (ruff+pyright), TypeScript (prettier+eslint+tsc), Go (gofmt+golangci-lint). Auto-fix and convergence loops.

a5c-ai/babysitter · 53 tokens

debug

Run /debug to find and fix a bug's root cause: a test failing for an unclear reason, /check verify finding a failure, or behavior being wrong. Runs a reproduce, localize, hypothesize, test, fix, verify loop, makes the minimal fix, and hands a regression test to /test. No features, no extra refactors.

jsmastery-pro/skills · 74 tokens

test-driven-bug-fix

Defines reproduce-red-green-refactor bug fixes. Load when correcting a defect with a regression test.

bostonaholic/team · 26 tokens

runtime-verification

Verify code works at runtime through build verification (mandatory), LSP diagnostics, ad-hoc verification for projects without frameworks, E2E and smoke tests, and visual verification (screenshot-analyze-verify for UI changes). Skip whitelist strictly enforced (markdown-only, config-only, dependency-bump-only with…

synaptiai/synapti-marketplace · 116 tokens

pdlc-fix

An automated bug-fixing workflow that locates a reported bug, reproduces it with a test, applies a minimal fix, runs tests, and updates the relevant documentation.

kanfu-panda/pdlc-skills · 25 tokens