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/emilvaintrub/showrunner/biblenpx skills add emilvaintrub/ShowRunner --skill biblegit clone --depth 1 https://github.com/emilvaintrub/ShowRunnerWrote 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/emilvaintrub/showrunner/bible)<a href="https://agentmods.dev/skills/emilvaintrub/showrunner/bible"><img src="https://agentmods.dev/badge/skills/emilvaintrub/showrunner/bible.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 | $0.00099 | $0.00579 |
| Opus 5 | $0.00049 | $0.00290 |
| Sonnet 5 | $0.00020 | $0.00116 |
| Haiku 4.5 | $0.00010 | $0.00058 |
Grade A, and why
bible 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 6d 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 — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Bible
Reconcile intent against reality and render one combined architecture and capabilities document. The Bible writes nothing new about the product: every claim traces to a repository path, interface, or named report.
Load
- Read
../core/method.md. - Read the active project's
.claude/showrunner/config.md. - Read method.md.
- Load
../core/merge.mdformerge. - Load templates/bible.md for
sync. - Load templates/merge.md for merge readiness.
Route
init: infer Bible source bindings (Forge spec, Arc ship reports, Sentry sweep, security, and pen-test outputs, and repository-structure hints for component, data-model, and interface extraction) and fill the Bible config section.sync: at an exact commit, extract components, the data model, and interfaces from the repository; reconcile against the Forge spec, Arc ship reports, and Sentry artifacts; apply the evidence rule and the intent-not-shipped drift gate; batch human-owned framing and ambiguous-evidence questions; then render templates/bible.md tobible.output.path.merge: after a cleansync, required smoke, and explicit approval, reuse the shared two-commit ceremony.
Hard Stops
- Never edit product code, configuration, schema, or risk memory.
- Never assert an architectural claim or a capability without a citing repository path, interface, or named report.
- Never list a spec-promised capability the repository does not implement as shipped; record it under Intent-Not-Shipped instead.
- Never render from a dirty worktree or a tip that does not match the exact commit being documented.
- Never restate Sentry's full threat model; Part II's security section is a summary with a pointer to the authoritative security document.
- Never choose audience framing, what counts as a "capability," or resolve ambiguous evidence on the human's behalf.
- Never merge without explicit human approval and the shared ceremony.
- Never import facts or wording from another project.
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.
- 6d ago First seen · 51 lines · 99 tokens per session scan A dd630cbd90bb
bible is a skill published in the GitHub repository emilvaintrub/ShowRunner (2 stars, last pushed 2mo ago), licensed MIT. It adds 99 tokens to every session and 579 once invoked, about $0.0005 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
kyzo-plan-write-story
Author or revise board story contracts (name, sources, condemned, ceiling, choice, allowlisted T#s, Final QA DoD). Not epic authorship (write-epic), not T# execution (run-story), not raw board moves (manage-board).
kyzo-plan-write-epic
Author or revise board epics — name + outcome as the value state the story group creates. Not story contracts (write-story), not T# runs (run-story), not column moves (manage-board).
kyzo-plan-run-story
Parent after startstory — arm paths, spawn demolition then each T#, run board Check, spawn judge, allowlist-commit on PASS / path-restore on FAIL, Final QA verdict then checkfinalqa. Not write-story or manage-board.
kyzo-plan-manage-board
ONLY board read/write — plan MCP tools, never raw gh. Tool descriptions are role×verb. Not write-story or run-story.
kyzo-plan-task-completion-request
Minimal completionrequest the development-task sends to the judge after allowlist edits. Not board writes, not git, not tree meters.
worker-integration
Worker-Agent integration for intelligent task dispatch and performance tracking.