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.
npx skills add infiniV/ultra-ml-intern --skill model-provenancegit clone --depth 1 https://github.com/infiniV/ultra-ml-internWrote 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/infiniv/ultra-ml-intern/model-provenance)<a href="https://agentmods.dev/skills/infiniv/ultra-ml-intern/model-provenance"><img src="https://agentmods.dev/badge/skills/infiniv/ultra-ml-intern/model-provenance/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/infiniv/ultra-ml-intern/model-provenance"><img src="https://agentmods.dev/badge/skills/infiniv/ultra-ml-intern/model-provenance.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00134 | $0.03581 |
| Opus 5 | $0.00067 | $0.01791 |
| Sonnet 5 | $0.00027 | $0.00716 |
| Haiku 4.5 | $0.00013 | $0.00358 |
Grade B, and why
model-provenance scanned grade B with 1 finding 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.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
cat ~/.claude/model-provenance/<slug>/SOURCES.md 2>/dev/null How it starts
The opening of the file, as written. The whole thing — 259 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Model Provenance
Given a model name, produce a local, verified, self-contained archive of its actual code and papers, then register a memory so all future work on that model reads the real source instead of guessing.
When NOT to use
- The user wants to run/train the model now, and an archive already exists →
use the
groundingskill, which reads this archive and checks the code against it. This skill builds the archive;groundingspends it. - A generic literature review with no specific model → use research/deep-research.
Output layout (the "safe folder")
Always archive to the global root ~/.claude/model-provenance/<model-slug>/,
never inside the current project. The archive is a machine-wide source of
truth shared by every project (the memory in step 8 points other repos at this
absolute path), so it must not live under any one project's working dir and must
not be committed to a project's git history. Do not ask for or accept a
per-project location; if the user wants a copy in their project, symlink it
after the fact. Expand ~ to the real $HOME in every path you write down.
Slugify the exact variant: DINOv3 → dinov3, SAM 2 → sam2.
~/.claude/model-provenance/<model-slug>/
├── code/ # full git clones — canonical repo first, key community repos
├── key_code/ # extracted train loop, model def, inference; + MANIFEST.md
├── papers/ # <slug>.pdf + <slug>.metadata.json (title/authors/abstract/bibtex)
├── hub/ # per official checkpoint: config/preprocessor/tokenizer/chat-template,
│ # model card, revision sha, license+gated status — metadata, never weights
├── SOURCES.md # provenance manifest: every repo+paper+checkpoint, commit pin, WHY canonical
└── notes.md # synthesis: architecture, recipe, I/O contract, variants, how to run
Reuse what's already there. Before doing any network work, check whether
~/.claude/model-provenance/<slug>/ already exists. If it does, treat the
existing archive as the starting point and only fill gaps — do not re-clone,
re-extract, or re-download artifacts that are already present and valid. Each
step below states its own skip condition. If the user explicitly asks to
refresh, delete the relevant subdir(s) first, then re-run those steps.
What ships with it
4 files 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.
- 10d ago First seen · 259 lines · 134 tokens per session scan B bf946934f9b6
model-provenance is a skill published in the GitHub repository infiniV/ultra-ml-intern (3 stars, last pushed 23d ago), licensed MIT. It adds 134 tokens to every session and 3,581 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
engineering-discipline
Behavioral guidelines to reduce common agent coding mistakes — thinking before coding, simplicity, surgical changes, goal-driven execution, and mandatory verification for data/BI work. Adapted from andrej-karpathy-skills (MIT), with a data/BI engineering layer added. Use this skill whenever writing, editing, or…
self-awareness
Behavioral calibration system that helps Claude catch its own failure patterns in real-time. Triggers on EVERY response to run self-checks for verbosity compensation, hallucination risk, missed frustration signals, sycophantic patterns, and constraint violations. Use this skill continuously—it modifies how Claude…
fable-method
Use when a task carries real uncertainty — multi-step builds, debugging where the first theory may be wrong, research or analysis with claims to verify, anything touching data/APIs/files not yet opened, or work that will be handed off. Also when work keeps failing or stalling, before declaring anything done, or when…
fable-status
Use when the user asks for the fable status, overlay health, the calibration record, or what the method has been tracking in this workspace. Read-only report; changes nothing.
agent-orchestration
Use to decide whether to delegate a task to a specialist agent, do it inline yourself, or skip — a cost/benefit policy over the project's scoped agents. Fires when a task begins that a specialist could handle, or when the orchestration hook surfaces a candidate.
clean-code-audio
Use when writing JUCE/audio DSP code — real-time-safety and DSP-specific idioms (parameter smoothing, block processing, branch-free loops, denormals, message↔audio-thread model). Complements clean-code-cpp.