agentic-sdlc
Get Started

Deterministic Runtime For Agentic Delivery

Scale AI-assisted software delivery without losing engineering control.

agentic-sdlc gives your team a local-first execution runtime where workflows are repeatable, policy-governed, resumable, and fully traceable.

Start in 60 seconds See execution flow
58+ Skills indexed
40+ Workflows available
121 Passing tests
  • Deterministic trace IDs and consistent step order
  • Policy and permission gates before risky commands
  • Resumable state after interruption or crash
  • Audit-ready timeline for review and compliance

Why teams choose it

Clear value for real engineering risk

Predictable Outcomes

Deterministic orchestration removes run-to-run drift so teams can trust execution results.

Operational Safety

Runtime policy enforces trust tiers, permission boundaries, and safe command behavior by default.

Faster Recovery

Resumable workflow state lets teams continue from the failure point instead of starting from zero.

Audit Confidence

Timeline traces make delivery reviews, incident analysis, and release sign-off explicit and verifiable.

How it works

From prompt to production-grade workflow execution

01

Author Workflow Intent

Define goals in markdown workflows, roles, and skills with explicit operating constraints.

02

Run Under Guardrails

Execute through policy gates for trust tier, command safety, and resource budgets.

03

Observe Every Transition

Track state, logs, and costs with deterministic trace IDs for each execution step.

04

Resume And Scale

Recover from interruptions and expand with domain packs as your delivery system matures.

Decision clarity

What changes when you move beyond ad-hoc agent scripts

Ad-hoc scripts

  • Inconsistent behavior across runs
  • Weak failure recovery
  • Low policy and audit visibility

agentic-sdlc runtime

  • Deterministic execution and traceability
  • Resumable state and controlled retries
  • Policy-enforced, review-friendly delivery

Runtime proof

Real workflow input/output visualized with Remotion

Demo Source

This video is rendered from a real workflow instance export: dev-create-showcase-video-1772167469615-59889.

  • Show structure: architecture layers and capability folders
  • Show CLI: command surface teams run in real delivery
  • Show use cases: AI delivery, security scan gating, and dev workflows
  • Show output success: landing page screenshot as final artifact
  • AI-assisted narration and subtitle timeline generated from trace data

Regenerate end-to-end with: cargo run -- --workflow-id dev/create-showcase-video --template dev/create-showcase-video_prompt --task "show full project value" then bash .agents/skills/dev/remotion_io_visualizer/scripts/export_trace_for_remotion.sh <instance_id> and bash .agents/skills/dev/remotion_io_visualizer/scripts/render_remotion_demo.sh <instance_id>.

Quick start

Run your first deterministic workflow

cargo run -- --workflow valid_flow.md
cargo run -- workflow doctor
cargo run -- workflow setup
cargo run -- workflow check
cargo run -- workflow trace <instance_id> --timeline

Build with AI agents and keep your SDLC deterministic.

Standardize one workflow, enforce policy guardrails, and scale with skill packs that keep quality high as velocity increases.

Inspect execution model Run from CLI