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 commands/tacoda/keystone/bootstrapgit clone --depth 1 https://github.com/tacoda/keystoneWhat 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 | $0.00006 | $0.01676 |
| Opus 5 | $0.00003 | $0.00838 |
| Sonnet 5 | $0.00001 | $0.00335 |
| Haiku 4.5 | $0.00001 | $0.00168 |
Grade A, and why
bootstrap 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 2d 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 — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.
bootstrap
One-time initial charter scaffold. Detect the project's stack, seed corpus/state/, scaffold idiom directories with code-grounded globs, inventory computational guides, classify sensors, and generate the globs index. Run once per project.
Activities
Every activity below produces a concrete file write. Detecting, narrating, or summarizing is not enough — call the edit primitive and land the change before moving on.
- Detect the stack. Inspect the repo (
package.json,go.mod,pyproject.toml,Gemfile,Cargo.toml,requirements.txt,build.gradle, etc.) and list every primary language, framework, and notable library. - Seed
.charter/corpus/state/CODEBASE_STATE.md. Propose an edit that replaces every template placeholder with real values:- Detected stacks
- Tool commands (lint / type-check / test / build / coverage) — actual commands, not placeholders
- Region map (top-level directories → which stacks they hold) — this becomes the source of truth for the globs in step 3
- CI platform (GitHub Actions / GitLab CI / etc.)
- Seed stack idioms with globs. For each detected stack, scaffold
.charter/corpus/idioms/<stack>/and the paired.charter/guides/idioms/<stack>/. The guides directory gets aREADME.mddescribing the stack, and each seeded guide's frontmatter declaresglobs:derived from the region map written in step 2 — the globs reflect real code paths, not invented patterns. Populate idiom files as patterns emerge; new files inherit the region's globs unless they cover a sub-tree (in which case narrow them). - Inventory computational guides. Record LSP, formatter, and editor enforcement (
.editorconfig, pre-commit hooks, etc.) under.charter/guides/computational/. Each entry recordsglobs:set to the paths the tool actually covers (read from its config file) — this lets the stack-drift sensor compare documented vs. effective configuration. - Classify sensors. For each sensor in
.charter/sensors/, propose an edit toCODEBASE_STATE.mdrecording whether this adapter can run it. Inferential sensors (review-functional,review-security,review-risk,review-deployment,spec-adherence) and computational sensors get separate sections. - Generate
.charter/corpus/state/GLOBS_INDEX.md. Walk every guide under.charter/guides/and.charter/policies/*/guides/, read each guide'sglobs:frontmatter, and write the reverse-index (glob pattern → list of guides claiming it). Touch only the## Indextable; preserve everything else. The index ships with an empty<pattern>placeholder row — bootstrap replaces it with real data. Pointer-style adapters (Claude Code, Codex, Aider) read this index to gate idiom loading on the touched-files set. - Project to per-adapter rule surfaces. For each guide with
globs:declared, produce the agent-specific projection:- Cursor (if
.cursor/rules/exists in the install) — write.cursor/rules/keystone-<topic>-<name>.mdcfor every guide that declaresglobs:. The<topic>-<name>slug comes from the guide's path under.charter/guides/(e.g.guides/idioms/typescript/hooks.md→keystone-idioms-typescript-hooks.mdc). The.mdcfrontmatter mirrors the guide'sglobs:; the body is a single-line pointer at the source guide. Guides withoutglobs:get no.mdc. - Pointer-style adapters (Claude Code, Codex, Aider, Cline, Continue, Goose, Pi) —
GLOBS_INDEX.mdfrom step 6 is the projection. No per-guide file is written; each adapter'sorientplaybook reads the index. _generic— skip. This adapter does not honorglobs:and falls back to topic defaults.
- Cursor (if
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.
- 2d ago First seen · 86 lines · 6 tokens per session scan A e556c866496c
bootstrap is a command published in the GitHub repository tacoda/keystone (44 stars, last pushed 2mo ago), licensed MIT. It adds 6 tokens to every session and 1,676 once invoked, about $0.0000 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
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
constitution
Create or update the project constitution from interactive or provided principle inputs.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.