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.
git clone --depth 1 https://github.com/chohra-med/expo_boilerplateWrote 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/commands/chohra-med/expo_boilerplate/learn)<a href="https://agentmods.dev/commands/chohra-med/expo_boilerplate/learn"><img src="https://agentmods.dev/badge/commands/chohra-med/expo_boilerplate/learn/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/commands/chohra-med/expo_boilerplate/learn"><img src="https://agentmods.dev/badge/commands/chohra-med/expo_boilerplate/learn.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 | $0.00000 | $0.00837 |
| Opus 5 | $0.00000 | $0.00418 |
| Sonnet 5 | $0.00000 | $0.00167 |
| Haiku 4.5 | $0.00000 | $0.00084 |
Grade A, and why
learn 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 9d 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 — 67 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Command: learn — the learning loop (feedback → rules)
The loop that makes the harness tighten. The verifier finds a failure;
learnturns that failure — or any human feedback — into an enforced rule so the next run can't repeat it. Feedback never evaporates. This is the "loop" pillar, reframed: not a cron schedule, a learning schedule.
When to run it
- A human corrected you ("no, the data only lives in the backend", "stop doing X").
- The verifier returned FAIL — capture the root cause before fixing, so it becomes a rule.
- A code-review comment, a bug postmortem, a "remember this for next time".
- End of a
build: harvest what was learned this feature.
Invocation
spec-harness learn "<the feedback / correction / lesson>" # one-shot
spec-harness learn --from .memory/80-feedback.md # drain the inbox
Or in Claude Code: /spec-harness learn and paste the feedback.
The loop (6 steps)
1 — CAPTURE
Append the raw feedback to .memory/80-feedback.md (dated inbox), verbatim, with the context
that triggered it (the file, the wrong behavior, what was expected). Never lose the raw signal.
2 — DISTILL
Turn the specific incident into ONE generalizable, imperative, testable rule. Bad: "don't assume tags map." Good: "Verify the data model against the real code before trusting a spec's data step — a spec can be wrong about where data lives."
3 — CLASSIFY (which layer does this rule belong to?)
| Kind of lesson | Goes to | Why |
|---|---|---|
| Hard constraint, never-again | AGENTS.md (the ratchet) |
only tightens; load-bearing |
| Stack-specific gotcha | ai_rules/rules/frequent_rules.md |
the per-stack most-violated list |
| Behavioral / process | ai_rules/rules/core.md |
how the agent works |
| Durable pattern / insight | .memory/70-knowledge.md |
reusable knowledge, not a guard |
4 — INJECT
Write the distilled rule into the chosen file, dated, appended (never rewrite history). The ratchet only grows. If a near-duplicate rule exists, sharpen it instead of adding a second.
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.
- 9d ago First seen · 67 lines · 0 tokens per session scan A 77a632ba4a88
learn is a command published in the GitHub repository chohra-med/expo_boilerplate (32 stars, last pushed 7d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 837 tokens. 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.
Other commands, from other repositories
arc-skill
You are scaffolding a new React Native (Expo) project using the arc-skill architecture.
arc-connect
Connect API layer, storage, state management, authentication, and i18n to an existing React Native Expo project. Use after arc-scaffold to add backend integration. Trigger when the user mentions adding API calls, auth flow, login, storage, state management, internationalization, backend connection, or data fetching to…
arc-scaffold
Scaffold a new React Native Expo project with folder structure, dependencies, TypeScript config, linting, and navigation shell. Use when starting a new mobile app from scratch, when the user says 'create a new app', 'start a project', 'set up React Native', 'init Expo app', or needs project boilerplate. Trigger this…
arc-ui
Add theme system, reusable components, screen layouts, and mobile UX design to a React Native Expo project. Use after arc-scaffold to build the visual layer. Trigger when the user wants to add a theme, dark mode, UI components, buttons, inputs, screen layouts, design system, or visual polish to their React Native app.
arc-audit
Audit a React Native Expo project for mobile UX issues, performance problems, and architecture violations. Use to check code quality against arc-skill best practices. Trigger when the user asks to review code, check for issues, wants a health check, mentions touch targets, accessibility, performance audit, or says 'is…
arc-feature
Add a complete new feature domain to an existing React Native Expo project — types, API module, React Query hooks, screen, and components in one go. Use when adding a new entity like products, orders, chat, users, recipes, or any CRUD resource. Trigger when the user says 'add a feature', 'create a new screen for X'…