BuildCLI-Agents
AGENTS.md3 starsChanged 11 days agoMITView on GitHub
Band skills — load only the one matching the work in front of you: - .codex/skills/service/SKILL.md - .codex/skills/interface/SKILL.md - .codex/skills/store/SKILL.md - .codex/skills/verify/SKILL.md - .codex/skills/delivery/SKILL.md Also available: - .codex/skills/code-standard/SKILL.md → implementation quality gates and validation evidence - .codex/skills/rig/SKILL.md → Codex hooks, band path map, and audit journal Startup behavior (required): 1. Run $survey first to create or refresh .buildcli/context.md. 2. Refresh it whenever the stack, architecture, integrations, or standards change. 3. Before any task, load only the band skill matching the…
We split this AGENTS.md into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 25
Run $survey first to create or refresh .buildcli/context.md.
- Distinctive: line 27
Before any task, load only the band skill matching the work: service, interface, store, verify, delivery.
≈ “Common” is an estimate: at least 1 in 5 of the 36,368 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Agent Instructions
<!-- buildcli:autoload:start -->Codex pipeline skills are generated from .codex/commands/*.md and invoked with $name:
- $survey → profile the repo into .buildcli/context.md
- $brief → requirements with acceptance criteria; moves .buildcli/active
- $shape → architecture and a phased plan (reads .buildcli/active)
- $worklist → atomic units with dependencies and band tags
- $build → execute the worklist band by band
- $audit → check the implementation against the brief
- $forge → create or refresh a project-specific skill
Band skills — load only the one matching the work in front of you:
- .codex/skills/service/SKILL.md
- .codex/skills/interface/SKILL.md
- .codex/skills/store/SKILL.md
- .codex/skills/verify/SKILL.md
- .codex/skills/delivery/SKILL.md
Also available:
- .codex/skills/code-standard/SKILL.md → implementation quality gates and validation evidence
- .codex/skills/rig/SKILL.md → Codex hooks, band path map, and audit journal
Startup behavior (required):
- Run
$surveyfirst to create or refresh.buildcli/context.md. - Refresh it whenever the stack, architecture, integrations, or standards change.
- Before any task, load only the band skill matching the work: service, interface, store, verify, delivery.
- Each band skill names the exact block of
.buildcli/context.mdto read. Read that block, nothing else. - Missing critical information → mark
NEEDS CLARIFICATIONand continue on safe defaults.
Pipeline order:
$survey→ profile the repo into .buildcli/context.md$brief→ requirements with acceptance criteria; moves .buildcli/active$shape→ architecture and a phased plan (reads .buildcli/active)$worklist→ atomic units with dependencies and band tags$build→ execute the worklist band by band$audit→ check the implementation against the brief$forge→ create or refresh a project-specific skill
Navigation:
$pulse→ pipeline snapshot: stage, unit counts, next step$focus→ move the active blueprint pointer$patch→ minimal defect fix; add --trace for a defect blueprint$rig→ configure Codex hooks and the audit journal
Runtime (always prefer it over reading state files by hand):
.buildcli/runtime/bcx band <name>— load exactly one context band.buildcli/runtime/bcx header— the shared header, without any band.buildcli/runtime/bcx active [path]— read or move the active blueprint pointer.buildcli/runtime/bcx next— units ready to start, grouped by band.buildcli/runtime/bcx graph— dependency graph, critical path, cycle report.buildcli/runtime/bcx claim|done|block <id>— unit state transitions.buildcli/runtime/bcx verify— run the project's test command.buildcli/runtime/bcx status --json— pipeline snapshot.buildcli/runtime/bcx resume— where the project left off, in one screen.buildcli/runtime/bcx doctor— validate context, graph, and configuration
Invoke it by that path, not as a bare bcx — it is project-local and not on PATH.
It needs python3. With $rig --enforce, Codex hooks call bcx gate to block raw
context reads and cross-band writes. Without hooks, the runtime still gives you
deterministic band extraction, real scheduling, and executable verification.
Shared state:
.buildcli/context.md— project context, split into [band:*] blocks.buildcli/active— path to the active blueprint directory.codex/hooks.json— Codex lifecycle hooks written by$rigblueprints/<kind>/<slug>/— brief.md, shape.md, worklist.md, audit.md
Multi-agent relay:
- brief + shape → an analysis-focused agent (Gemini, Claude)
- build → Codex is strongest here: one band, one unit, focused generation
- audit → Gemini for gap analysis, Codex for coverage checks
Bootstrap command:
./buildcli/scripts/bootstrap.sh --repo . --agent all --mode copy
<!-- buildcli:autoload:end -->Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 28,693 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository19 other files it gives its agents.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

