th08: Skill for Codex

.agents/skills/th08-matching/SKILL.md

th08-matching is a skill for Codex from N0zoM1z0/th08. It costs 62 tokens per session (7,431 once invoked), scanned A, original, MIT.

A focused build-and-comparison workflow for TH08, a specific software target, using the repository's VC7 compiler setup. It checks whether generated object-code functions match an attested version 1.00d target.

In plain words
What is it for?
Use it to verify target functions, build selected object files, inspect relocations, diagnose code-generation differences, and compare results against the configured target.
Why use it?
It helps locate compiler, relocation, or binary differences when exact output matters, while requiring evidence about the target and application binary interface before changes are made.

Skill for Codex

Written for Codex: agents/openai.yaml present. Also seen: installed under .agents/ (shared by several agents); $skill-name invocation.

This is N0zoM1z0/th08's own configuration. It tells Codex how to work on th08 itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything th08 configures →

Needs its repository: it runs a file that does not travel with it, so clone the repository first. The line is python3 scripts/verify-target.py.

Reuse

Borrowing it

Nothing to install: this file belongs to N0zoM1z0/th08. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/N0zoM1z0/th08/main/.agents/skills/th08-matching/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/N0zoM1z0/th08

Made for: Codex.

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.

agentmods badge for th08-matching

README.md
[![agentmods](https://agentmods.dev/badge/skills/n0zom1z0/th08/th08-matching/github.svg)](https://agentmods.dev/skills/n0zom1z0/th08/th08-matching)
Your own site
<a href="https://agentmods.dev/skills/n0zom1z0/th08/th08-matching"><img src="https://agentmods.dev/badge/skills/n0zom1z0/th08/th08-matching/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.

agentmods 80×15 button for th08-matching

Your own site · 80×15
<a href="https://agentmods.dev/skills/n0zom1z0/th08/th08-matching"><img src="https://agentmods.dev/badge/skills/n0zom1z0/th08/th08-matching.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 62 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 7,431 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. Third-party audits
  • NVIDIA SkillSpector pass 7 Sept 2026
How audits are shown
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 5.1 $0.00062 $0.07431
Opus 5 $0.00031 $0.03716
Sonnet 5 $0.00012 $0.01486
Haiku 4.5 $0.00006 $0.00743

Measured 10d ago against content hash 805e90eb81e1, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

th08-matching 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.

.agents/skills/th08-matching/SKILL.md · 379 lines

How it starts

The opening of the file, as written. The whole thing — 379 lines — stays where its author put it; the contents beside it link to each section on GitHub.

TH08 exact matching

Use with $th08-re only after the target boundary, semantics, and ABI are supported by evidence.

Canonical loop

  1. Verify the target and list configured units:

    python3 scripts/verify-target.py
    python3 scripts/compare-function.py --list
    
  2. Read the selected [[units]] entry in config/match-units.toml. Confirm its target address/size, decorated COFF symbol, object path, and every relocation against config/mapping.csv, config/reccmp-*.csv, and target disassembly. If an original delinked object is required and Ghidra is configured, generate it through the verified disposable import:

    python3 scripts/export_ghidra_objs.py --import-csv
    
  3. Build the entry's object basename. For an object recorded as build/ScreenEffect.obj, run:

    python3 scripts/build.py --build-type=objdiffbuild --object-name ScreenEffect.obj
    

    If VC7 emits the right implicit constructor bytes but omits a COFF function- definition auxiliary record, declare and define the natural empty constructor explicitly, rebuild, and compare again. Do not synthesize behavior; member construction must remain compiler-driven.

  4. Compare the configured function:

    python3 scripts/compare-function.py UNIT --json
    
  5. Treat exact, mismatch, and error literally. Classify a mismatch before editing source: function boundary, ABI/layout, compiler flags, decorated symbol, REL32/DIR32 mapping, source/control-flow order, or translation-unit effect.

  6. Change one source-plausible expression or declaration, rebuild the same object, and rerun the same unit.

For a read-only search of already built, source-emitted leaves, use the conservative discovery gate:

nice -n 15 python3 scripts/analysis/propose-exact-units.py --min-size 0x80

It emits review-only candidates only when the target-pinned isolated extent, full relocation replay, direct-call multiset, named destination ledgers, and unique object ownership agree. --allow-unimplemented may reveal source-built functions missing from implemented.csv; it never updates tracking. Inspect the source and target, add a reviewed unit, and obtain this skill's canonical compare-function.py exact result before entering either ledger.

Read the full file on GitHub · 379 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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. 10d ago First seen · 379 lines · 62 tokens per session scan A 805e90eb81e1

Subscribe to this mod's changes

th08-matching is a skill published in the GitHub repository N0zoM1z0/th08 (162 stars, last pushed 4d ago), licensed MIT. It adds 62 tokens to every session and 7,431 once invoked, about $0.0003 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 2026-08-30.

Related

Other skills, from other repositories

idapython

IDA Pro Python scripting for reverse engineering. Use when writing IDAPython scripts, analyzing binaries, working with IDA's API for disassembly, decompilation (Hex-Rays), type systems, cross-references, functions, segments, or any IDA database manipulation. Covers ida modules (50+), idautils iterators, and common…

mrexodia/ida-pro-mcp · 77 tokens

ghidra-agent-cli

Rust CLI reference for the headless-ghidra pipeline. Covers command syntax, flags, output contract, artifact paths, and workspace layout for all ghidra-agent-cli subcommands. Load when: constructing a ghidra-agent-cli command, interpreting its output, resolving a flag or artifact path question, or debugging CLI…

ByteLandTechnology/headless-ghidra · 95 tokens

headless-ghidra-analyze-function

Thoroughly analyze a single function by first recovering its type definitions, constant definitions, vtables, function name and signature, then producing the final decompilation. Use when the user wants a complete per-function analysis following the strict recovery order: types → constants → vtables → function…

ByteLandTechnology/headless-ghidra · 72 tokens

headless-ghidra

Entry skill for the Headless Ghidra YAML-first reverse-engineering pipeline. Use when the user asks to analyze, decompile, triage, resume, or iterate on a binary target with Ghidra/headless-ghidra. Reads artifacts/ /pipeline-state.yaml, routes P0–P4 phase skills, runs gate checks, and manages review pauses. Performs…

ByteLandTechnology/headless-ghidra · 87 tokens

headless-ghidra-baseline

P1 phase skill for Headless Ghidra baseline and runtime evidence. Use after P0 when the target must be imported into Ghidra, auto-analyzed, exported to baseline YAML, and given reproducible runtime or hotpath observations without decompiling function bodies.

ByteLandTechnology/headless-ghidra · 63 tokens

headless-ghidra-batch-decompile

P4 phase skill for Headless Ghidra selected function substitution. Use after P3 when an approved batch of functions should have metadata applied, be decompiled through Ghidra, and be recorded as per-function capture/substitution YAML.

ByteLandTechnology/headless-ghidra · 58 tokens