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 agentmods add skills/donald-ada/workinggenius/decision-recordnpx skills add donald-ada/workinggenius --skill decision-recordgit clone --depth 1 https://github.com/donald-ada/workinggeniusWrote 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/donald-ada/workinggenius/decision-record)<a href="https://agentmods.dev/skills/donald-ada/workinggenius/decision-record"><img src="https://agentmods.dev/badge/skills/donald-ada/workinggenius/decision-record.svg" alt="Measured on agentmods" 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 | $0.00073 | $0.00721 |
| Opus 5 | $0.00036 | $0.00360 |
| Sonnet 5 | $0.00015 | $0.00144 |
| Haiku 4.5 | $0.00007 | $0.00072 |
Grade A, and why
decision-record 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 yesterday.
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 — 32 lines — stays where its author put it; the contents beside it link to each section on GitHub.
The Decision Index
A decision is never written twice. The flow already writes every decision at full strength — the kill-reasons, the survivor's fight, the user's confirmation — into the work file where it was made. What those records lack is a door: a fresh session judging a new design cannot trawl every done file for the one decision it is about to contradict. The door is .genius/DECIDED.md, beside the work files wherever the project pins them — one line per settled decision, each pointing at the fight that settled it. The verdict is the index line; the fight stays where it was written.
The line
- **Raw SQL over any ORM** — schema drift burned v1; queries are few and
hand-debugged. [The fight](checkout-rework/checkout-rework.log.md#discernment)
Verdict in bold, the why in one line, the link to the record that holds the whole battle. The link is relative to .genius/, where DECIDED.md sits, so it carries the work's folder, as above. A decision that landed outside the flow — no work file, no battlefield — carries its why in the line itself: there the line is the record, so give the why one honest sentence rather than a verdict alone.
The test
An index line is earned by one question: would a future stranger re-fight this? — propose the killed thing again, redo the settled study, "fix" the deliberate deviation. It is the same re-litigation test the kill-reason already answers inside one piece of work, asked across all future work. Most decisions fail it — they bind only their own work, and their work file holds them fine. The ones that pass: architectural shape, boundary no-s, deliberate deviations from the obvious path, constraints invisible in the code, rejections that will be re-proposed by someone who wasn't there.
The discipline
- Written when the record is finished — Tenacity's close-out asks the test of everything the work settled; a standalone
/architectstudy writes its line when the design is confirmed. Never a second write-up: the line points, the record carries. - Read before contradicting. The index is the ammunition
/discernand/architectload: a design that touches a settled decision either dies of the fight on record or names the decision it overturns and why. - Overturning moves the line. A new fight that beats a settled decision rewrites its line — new verdict, new link — and the old record stays exactly as written, because a record's worth is that it is what was written then. The index carries what the project believes now; the files carry how it came to believe it, including everything it used to.
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.
- yesterday Changed 8ea2b15d14d7
- 6d ago First seen · 32 lines · 73 tokens per session scan A 2198a8c6a9e5
decision-record is a skill published in the GitHub repository donald-ada/workinggenius (6 stars, last pushed yesterday), licensed MIT. It adds 73 tokens to every session and 721 once invoked, about $0.0004 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.
Other skills, from other repositories
self-improve
Self-improve this skills repo — surface and (safely) apply quality improvements across four parallel read-only lenses (Clarity, Leanness, Coverage, Ergonomics). Sibling of /self-audit — audit fixes rule violations, improve raises quality where no rule is broken. Local meta-skill, independent of the /task: pipeline.
self-audit
Self-audit this skills repo against CLAUDE.md invariants, the artifact contract, and README/CLAUDE.md/docs sync via three parallel read-only subagents. Local meta-skill — independent of the /task: pipeline.
grill
Interrogate a plan or decision one question at a time before capture, keeping a decision-plus-rationale ledger, then route to the right capture skill.
dw-doctor
Read-only environment diagnostic for a solo-lane repo: whether the tools the hooks and skills assume are installed and whether the repo's guardrails will really fire, each gap reported with a copy-paste fix.
dw-handoff
Compact the live session into .ai/work/ /HANDOFF.md — how far into the current task you are, what is applied but uncommitted, and which dead ends are already ruled out, so the next context window resumes mid-task. Explicit-invoke only.
dw-next
The solo lane's build step and its resume point in one skill: report where the active .ai/work/ /CHANGE.md stands — read from disk, so it survives a /clear — then build every remaining unticked task, one commit each.