Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add paulbaranowski/wild-horses/plugin install pyrightWrote 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/paulbaranowski/wild-horses/run-and-fix)<a href="https://agentmods.dev/commands/paulbaranowski/wild-horses/run-and-fix"><img src="https://agentmods.dev/badge/commands/paulbaranowski/wild-horses/run-and-fix/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/paulbaranowski/wild-horses/run-and-fix"><img src="https://agentmods.dev/badge/commands/paulbaranowski/wild-horses/run-and-fix.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 | $0.00065 | $0.06762 |
| Opus 5 | $0.00032 | $0.03381 |
| Sonnet 5 | $0.00013 | $0.01352 |
| Haiku 4.5 | $0.00006 | $0.00676 |
Grade A, and why
run-and-fix 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 11d 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 — 376 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Pyright: Run and Fix
Run pyright on the Python code in this project and fix the errors it finds, following the rule-specific patterns in the bundled playbook.
Pattern catalog. The playbook is split across five files at ${CLAUDE_PLUGIN_ROOT}:
reference.md— index, setup, triage, suppression policy, narrowing artifacts, documented-API preference, CI, parallel dispatch, external-finding verification, editor-autofix warning, serialization compat, config-intent principle, assert-vs-raise, consolidation pass (orchestrator-level).rules.md— fix recipes keyed on pyright rule name (reportOptionalMemberAccess,reportArgumentType,reportTypedDictNotRequiredAccess, …).libraries.md— library-stub workarounds (bitstring, scipy, tornado, matplotlib, Beanie, Supabase, litellm, pydantic, PIL, tenacity, pymongo, …).bugs.md— signals that pyright has uncovered a real bug (to flag for the user, not silence).suggestions.md— Phase 5 "Suggested improvements" procedure: signals to combine, sort order, save flow.
Read reference.md at the start of the run for the index and the triage/policy framework. Re-consult rules.md and libraries.md when triage surfaces specific rules or library names. If ${CLAUDE_PLUGIN_ROOT} isn't substituted in this context, find the files with Glob "**/pyright/reference.md" and read the siblings alongside it.
Arguments: "$ARGUMENTS"
Phase 1 — Parse args and verify setup
Step 1: parse $ARGUMENTS
Expected tokens, all optional:
- Positional
level∈ {basic,standard,strict}. Overrides the config'stypeCheckingModefor this invocation. --persist— if a level is given and the run reaches zero errors, write that level back to config.--ratchet— climbbasic→standard→strict, fixing to zero at each rung. Mutually exclusive with an explicitlevel; implies--persistat each rung.--scope <path>— restrict pyright invocation and fix work to a subpath. Default: the config'sinclude.--intent <silence|improve|bugs-only>— declares the lean Phase 3 should take when fixing errors. See Phase 3 § "Apply the chosen intent" for the full, authoritative definitions. If omitted, the command prompts in Phase 2 after showing triage.--no-suggestions— skip the "Suggested improvements" section in Phase 5. Default: include it.
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.
- 11d ago First seen · 376 lines · 65 tokens per session scan A f88c65fef8f0
run-and-fix is a command published in the GitHub repository paulbaranowski/wild-horses (12 stars, last pushed 23d ago), licensed MIT. It adds 65 tokens to every session and 6,762 once invoked, about $0.0003 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 2026-08-30.
Other commands, from other repositories
build-fix
Incrementally fix Python type and lint errors.
async-check
Vérification Patterns Async Python.
type-coverage
Vérification Couverture des Types Python.
lint:explain
(Ruff) Explain a ruff error and show the PROPER solution (without noqa!).
typecheck:explain
(Typing) Explain type error and show the correct solution (without type: ignore).
typecheck
(Typing) Run mypy/pyright for type checking.