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 cboone/agent-harness-plugins --skill write-formalization-roadmapgit clone --depth 1 https://github.com/cboone/agent-harness-pluginsWrote 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/cboone/agent-harness-plugins/write-formalization-roadmap)<a href="https://agentmods.dev/skills/cboone/agent-harness-plugins/write-formalization-roadmap"><img src="https://agentmods.dev/badge/skills/cboone/agent-harness-plugins/write-formalization-roadmap/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/cboone/agent-harness-plugins/write-formalization-roadmap"><img src="https://agentmods.dev/badge/skills/cboone/agent-harness-plugins/write-formalization-roadmap.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 | $0.00049 | $0.01096 |
| Opus 5.5 | $0.00020 | $0.00438 |
| Sonnet 5.5 | $0.00010 | $0.00219 |
| Haiku 4.5 | $0.00005 | $0.00110 |
Grade A, and why
write-formalization-roadmap 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 8d 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 — 105 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Write Formalization Roadmap
Apply the structural conventions from the reference files below when drafting or editing a multi-milestone formalization roadmap.
When to Use
A formalization roadmap is the long-lived planning document for a
multi-milestone mechanization of a specific paper, theorem, or related
cluster of results. It lives under docs/plans/todo/ until the whole
program ships, then moves to docs/plans/done/. It coordinates work
that will span many sessions, many commits, and usually multiple
worktrees or branches.
Use this skill when the planning document is a roadmap in that sense. Do not use it for:
- Single-bounded implementation plans (one feature, one fix, one
reorganization). Those follow the ordinary
docs/plans/todo/plan conventions, not the roadmap schema. - Paper manuscripts, framework notes, or prose about mathematics. Use
write-math,write-pandoc-markdown, andwrite-latexinstead. - Per-milestone plan files that a roadmap spins out. Those are ordinary implementation plans; the roadmap's Section 6 entry is the short summary, and the spun-out file is the elaboration.
Core Principles
- Structure is the contract. The 10-section schema and the 5-part milestone anatomy are the interface between the roadmap author and every future session that reads it. Deviating silently makes the roadmap harder to pick up cold; deviating with an explicit justification recorded in the roadmap is fine.
- Resolutions freeze in-doc. Once a scope question is answered, record the resolution in Section 4 rather than leaving it open or scattering the answer across chat history. Later sessions must be able to reconstruct the decision from the document alone.
- Tests are part of the milestone. Every proof milestone has a named test module that lands in the same checkpoint. Milestones without named tests are not complete milestones.
- Checkpoint gates are buildable, linted, and testable. There is no "partial milestone" state. Either the whole checkpoint compiles and passes its gates or the milestone slips; record the slip explicitly.
- Cite equations, not gestures. Paper citations must name specific equation numbers, theorem numbers, and page numbers from the verified transcription. "See Theorem 3" is not a citation.
- Roadmap entries summarize; plan files elaborate. A Section 6
milestone entry is a short, fixed-shape summary. The full
elaboration lives in a spun-out plan file under
docs/plans/todo/<date>-<milestone-slug>.md. Do not duplicate.
What ships with it
6 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.
- 8d ago Changed · +1 lines · +18 tokens per session 08fe7b0a5cf1
- 15d ago Changed · -8 tokens per session e1cf180fbb07
- 26d ago Changed · -13 lines · -170 tokens per session 0cc61157f1d1
- 1mo ago Changed bb4efc1da942
- 1mo ago First seen · 117 lines · 209 tokens per session scan A 75337435515f
write-formalization-roadmap is a skill published in the GitHub repository cboone/agent-harness-plugins (2 stars, last pushed yesterday), licensed MIT. It adds 49 tokens to every session and 1,096 once invoked, about $0.0002 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 2026-09-03.
Other skills, from other repositories
hypothesis-generation
Formulate evidence-bounded scientific questions, candidate hypotheses, rival explanations, causal or associational claims, discriminating predictions, measurements, and preregistration-ready analysis plans. Use when turning observations or preliminary findings into transparent, testable research plans without treating…
datamol
Pythonic wrapper around RDKit with simplified interface and sensible defaults. Preferred for standard drug discovery including SMILES parsing, standardization, descriptors, fingerprints, clustering, 3D conformers, parallel processing. Returns native rdkit.Chem.Mol objects. For advanced control or custom parameters…
reference
Optional method and tool reference for the Research Director -- stage playbooks, the research-loop and episode mechanics, tool routing, statistical execution and figures, and domain packs. Advice to consult when useful, not a required order.
direct-study
Drive a scientific study with data-first experiments and hypothesis revision, parallel literature, evidence-supported analysis and conclusions, then an appropriate manuscript, using phases as evidence milestones while preserving exact source, run, artifact, decision, analysis-plan, manuscript, and journal-validation…
analyze
Run deterministic local descriptive, explicit probability-distribution fitting, inferential, nonparametric, regression, bounded Gaussian random-intercept mixed-model, two-factor response-surface, survival, multivariate PCA, time-series diagnostic, contingency, correction, and confidence-interval analyses with…
deextinction-feasibility
Decompose a dinosaur revival question into independently evidenced stages with explicit stop conditions and Research Director review.