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.
curl -O https://raw.githubusercontent.com/N0zoM1z0/th08/main/.agents/skills/th08-semantic/SKILL.mdgit clone --depth 1 https://github.com/N0zoM1z0/th08Wrote 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.
[](https://agentmods.dev/skills/n0zom1z0/th08/th08-semantic)<a href="https://agentmods.dev/skills/n0zom1z0/th08/th08-semantic"><img src="https://agentmods.dev/badge/skills/n0zom1z0/th08/th08-semantic/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.
<a href="https://agentmods.dev/skills/n0zom1z0/th08/th08-semantic"><img src="https://agentmods.dev/badge/skills/n0zom1z0/th08/th08-semantic.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00065 | $0.01971 |
| Opus 5.5 | $0.00026 | $0.00788 |
| Sonnet 5.5 | $0.00013 | $0.00394 |
| Haiku 4.5 | $0.00006 | $0.00197 |
Grade A, and why
th08-semantic 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 162 lines — stays where its author put it; the contents beside it link to each section on GitHub.
TH08 semantic reconstruction
Work only from the verified Japanese TH08 1.00d target and the current authored
source. Read AGENTS.md, docs/ARCHITECTURE.md, docs/RE_WORKFLOW.md,
docs/RE_HANDOFF.md, docs/SEMANTIC_RECONSTRUCTION.md, and
docs/SEMANTIC_PLAYBOOK.md before editing.
Preflight
Run the target/tracking preflight and keep config/claims.csv header-only. Use
scripts/analysis/report-semantic-debt.py as a candidate router, never as a
completion metric. Select one structure and one coherent field family.
Confirm every touched function in config/mapping.csv,
config/reccmp-functions.csv, and config/match-units.toml. Establish the
current accepted-unit baseline for every affected object before changing a
shared declaration.
Recover meaning
Record target facts, corroboration, inference, and unknowns separately. Inspect
all reads and writes, widths, signedness, bit operations, callers, strings, and
canonical object ownership. TH06/TH07 and upstream names are corroboration
only when TH08 evidence agrees. IDA evidence remains blocked until the active
database passes docs/IDA_MCP.md.
Treat ownership as four separate questions: storage owner, semantic owner, translation-unit/COMDAT emission owner, and production/probe/portable build owner. Evidence for one does not establish the others. Close the relevant producer, consumer, and state-transition set before changing any of them.
Prefer real fields, aggregates, enums, bitfields, and member/index expressions.
Preserve opaque storage when only its extent is known. Add focused C_ASSERT
checks for relied-on sizes and offsets, remembering that layout assertions do
not prove names.
When one fixed storage range has several target-proven roles selected by a
discriminator, a translation-unit-local typed union overlay may be safer than
changing a widely included VC7 header. Anchor the single cast boundary to a
real member with offsetof/sizeof assertions, keep generic initialization
explicit, and use role-specific members everywhere else. This exception is
for evidenced variant storage, not a way to assign meaning to unknown bytes.
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.
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.
- 1mo ago Changed · +6 lines cf77233886f0
- 2mo ago Changed · +62 lines · +3 tokens per session d8ec983b4054
- 2mo ago First seen · 94 lines · 62 tokens per session scan A fa31be82ebc4
th08-semantic is a skill published in the GitHub repository N0zoM1z0/th08 (55 stars, last pushed yesterday), licensed MIT. It adds 65 tokens to every session and 1,971 once invoked, about $0.0003 per session on Opus 5.5. 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.
th08-semantic compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| th08-semanticN0zoM1z0/th08 | Skill | 55 | yesterday | MIT | A | 65 | Codex |
| zoom-meeting-sdk-unrealanthropics/knowledge-work-plugins | Skill | 35k | 2d ago | Apache-2.0 | A | 41 | Claude Code |
| unreal-cpp-gameplaygamedev-skills/awesome-gamedev-agent-skills | Skill | 6.9k | 10d ago | Apache-2.0 | A | 105 | Claude Code |
| idapythonmrexodia/ida-pro-mcp | Skill | 13k | 1mo ago | MIT | A | 77 | Claude Code |
| llvm-learninggmh5225/awesome-llvm-security | Skill | 890 | 1mo ago | MIT | A | 39 | Claude Code |
| mlir-developmentgmh5225/awesome-llvm-security | Skill | 890 | 1mo ago | MIT | A | 54 | Claude Code |
| compiler-developmentgmh5225/awesome-llvm-security | Skill | 890 | 1mo ago | MIT | A | 46 | Claude Code |