build
Skill3 starsChanged 11 days agoMITView on GitHub
Execute a worklist band by band, scheduled by the runtime, with live progress and sub-agent fan-out.
We split this skill into the rules it gives an agent
6 rules. 0 are common, 4 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 137 edited versions we read keep these.
4
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 30
Run .buildcli/runtime/bcx graph.
- Distinctive: line 32
Ask the runtime what can start:
- Distinctive: line 55
Run the suite and record the result:
- Very specific: line 114
Never read .buildcli/context.md.
- Distinctive: line 115
Never resolve a cross-band ripple inside the unit that found it.
- Distinctive: line 116
Update TodoWrite and the runtime together — a unit marked done in one and not the other is a lie.
“Common” is measured: kept by at least 6 in 10 of the 137 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Build
Arguments
(optional) Path to a blueprint directory. Omitted → the runtime reads the active pointer.
Output
Code changes + a build report.
Resolving the blueprint
The runtime resolves it. Do not read the pointer file yourself.
.buildcli/runtime/bcx active # the active blueprint, or an error telling you to run /brief
.buildcli/runtime/bcx graph # units, critical path, and any structural problem.buildcli/runtime/bcx graph exits non-zero when the worklist is not schedulable — a dependency cycle, an
unknown band tag, a unit with no check. Fix those before building anything; a cycle means the
plan is wrong, not the code.
Steps
- Run
.buildcli/runtime/bcx graph. Non-zero exit → report the problems and stop. - Register every unit in TodoWrite as
pendingbefore touching a file. - Ask the runtime what can start:
It returns the ready units grouped by band — unblocked, dependencies satisfied. That grouping is the fan-out plan; you do not compute the order yourself..buildcli/runtime/bcx next --json - For each unit:
a. Flip the unit to.buildcli/runtime/bcx claim W03 # marks it in progress; the write gate now scopes to its band .buildcli/runtime/bcx band service # load exactly the band that unit belongs toin_progressin TodoWrite. b. Load.claude/skills/<band>/SKILL.md. c. Write the smallest change that satisfies the unit's check. d. Close it out:
e. Anything that ripples into another band becomes a follow-up unit, never an in-place fix. With the harness enforced, the write gate will block you from touching another band's files while this unit is claimed. That block is the design working, not an obstacle..buildcli/runtime/bcx done W03 # or: .buildcli/runtime/bcx block W03 --reason "<what stopped it>" - Repeat from step 3 until
.buildcli/runtime/bcx nextreports nothing ready. - Run the suite and record the result:
.buildcli/runtime/bcx verify - Emit the report.
Sub-agent fan-out
.buildcli/runtime/bcx next --json returns more than one band → those groups are independent and can run
concurrently. Spawn one sub-agent per band and hand each:
- its unit list, from the runtime's output
- its own identity, and the instruction to pass
--agent <id>on every runtime command - the path to its band skill,
.claude/skills/<band>/SKILL.md - the instruction to load context with
.buildcli/runtime/bcx band <band>and nothing else brief.md, for reference
Enforcement survives the fan-out only if each worker is distinguishable. The gate resolves the
caller's own claim and holds it to that band, so named workers each stay inside their own. Give every
worker a distinct --agent, and export a matching BCX_AGENT_ID in its environment — the gate runs
as a hook subprocess and reads the environment, since a hook event carries no agent field.
Where the host cannot give each worker its own environment, they all look like one caller, the band
is ambiguous again, and the gate allows everything. That is not a new failure — it is how the gate
behaved before identities existed — but it is invisible unless you look. .buildcli/runtime/bcx doctor
names it: two live claims sharing one owner is reported precisely because enforcement does not error
in that state, it silently stops. Check it once at the start of a fan-out.
Bands ready in parallel are not the same as files that do not collide. Two bands can own edits to the
same file; .buildcli/bands.json is what makes the separation real, and a project without one gets no
path enforcement at all. Fan out on disjoint files, not just disjoint bands.
Merge the sub-agent results before reporting.
Output
## Build Report — <feature>
### Landed
- [band] W01: what it does → files changed
### Blocked
- W05: reason
### Cross-band follow-ups
- [service → interface] what needs attention
### Verification
- command: npm test
- result: PASS (exit 0)
### Quality gates
- Gate name: Pass | FailRules
- Let the runtime schedule. Never hand-derive the dependency order from the markdown.
- Never read
.buildcli/context.md. One band, through.buildcli/runtime/bcx band. - Never resolve a cross-band ripple inside the unit that found it.
- Update TodoWrite and the runtime together — a unit marked done in one and not the other is a lie.
- One concern per unit — no feature, fix, and refactor bundled into a single step.
Other versions149 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 149
- FalkorDB/FalkorDB★ 11ksame name
- microsoft/microsoft-ui-xaml★ 8.6ksame name
- NexaAI/nexa-sdk★ 8.5ksame name
- hathach/tinyusb★ 7.3ksame name
- parcadei/Continuous-Claude★ 3.9ksame name
- parcadei/Continuous-Claude-v2★ 3.9ksame name
- parcadei/Continuous-Claude-v3★ 3.9ksame name
- pytorch/TensorRT★ 3ksame name
- oxsecurity/megalinter★ 2.6ksame name
- iPlug2/iPlug2★ 2.4ksame name
139 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
Used alongsideProjects with it also use backprop, finances and 9 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
JuliusBrusseeSkillbackprop
Bug → spec protocol. When a bug is found or a test fails, trace the cause, decide whether a new §V invariant would catch recurrence, append to §B. This is the one non-obvious thing SDD does that plan-then-execute doesn't. Triggers on test failure, bug report, post-mortem, or explicit user ask.
Also in 4 projects that use this
GerardoRdz96Skillfinances
Use when Gera talks about his money — logging an expense ("I spent X on Y"), checking where he stands this pay-period, closing a period, planning debt payoff, onboarding his financial picture, or asking for budgeting/finance opinions. Servy's financial brain. Triggers — "I spent", "log this expense", "where am I this period", "close the period", "how's my budget", "debt plan", "should I pay off", "finances onboard", "/finances".
Also in 3 projects that use this
cpiprintSkillfavicon
Generate favicons from a source image
Also in 3 projects that use this
madeehahassanSkillGo-Live
Pre-launch checklist. Use before making any app public, sharing a link externally, or announcing a product launch. Adapts to Claude Code, claude.ai, and Cowork.
Also in 3 projects that use this
TrevokeSkilldefine
Use when a new feature request, bug report, or work request needs requirements defined before implementation
Also in 3 projects that use this
synaptiaiSkillgate
Create a BDSK review_gate artifact. Defines a checkpoint that controls whether generated output may proceed toward acceptance. Use when defining quality, security, or scope gates for execution plans.
Also in 5 projects that use this
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.
Reviews can't be read right now.
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.

