Dag Build
Decomposable coding task as a DAG: conductor plans, agent leaves build files concurrently, a deterministic oracle command gates, a heal fixpoint retries with failure context, halt-judge stops unproductive loops, checkpoints make runs resumable. Trigger: build this module across N files / 实装这个可分解任务 / parallel implementation with a test gate / dag-build. Skip: small single-file change (edit directl…
Install
$ agentstack add skill-abysscn-oh-my-dag-dag-build ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Links to your public security report.[](https://agentstack.voostack.com/security/report/skill-abysscn-oh-my-dag-dag-build) Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
/dag-build — parallel build with a hard oracle
For coding tasks that decompose into independently-buildable pieces and have a deterministic pass/fail command. The conductor plans the DAG, agent leaves edit files concurrently, the oracle command gates the result; on red, a heal round re-plans with the failure output injected; a halt-judge kills unproductive loops; node checkpoints make interrupted runs resumable.
Usage
bun run dag-build "" \
[--cwd ] # where the build happens
[--oracle-cmd "bun run typecheck"] # deterministic gate; exit 0 = pass
[--context ""|--context-file p] # spec/contract fed to the conductor
[--test | --test-glob ""] # run tests as part of the oracle
[--conductor-model M] [--agent-model M] [--leaf-model M]
[--agent-thinking off|low|medium|high|xhigh] [--leaf-ponytail]
[--max-fanout 8] [--heal N] # heal rounds on red oracle (default 1)
[--review|--no-review] [--review-gate G2|G3] # embedded dag-review before finish
[--resume ] # resume from checkpoints
[--allow-dirty] [--no-judge] [--judge-cap N]Exit code = oracle ∧ judge — safe to chain in CI.
Discipline
- No oracle, no dag-build. If you can't state a deterministic pass command, the
task isn't ready for parallel dispatch — tighten the contract first.
- Requires a clean git tree by default (
--allow-dirtyto override) so a failed run
is trivially revertible.
- Small, already-loaded single-file edits are cheaper done directly — dispatch has
fixed overhead; use this where width pays.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: AbyssCN
- Source: AbyssCN/oh-my-dag
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.