agentleFS
Sign inSign up

safe-agentic-workflow / rules

bybren-llc/safe-agentic-workflow/.cursor/rules/03-safe-ai-dlc.mdc

SAFe x AI-DLC fusion for running programs: in programs that adopt it, Bolts take the place of sprints and Units of Work of Features, plus Mob Elaboration and the human-in-the-loop validation gate. Opt-in per program. Activate when planning or running multi-issue programs.

Cursor rule405 starsChanged 2 months ago
---
description: "SAFe x AI-DLC fusion for running programs: in programs that adopt it, Bolts take the place of sprints and Units of Work of Features, plus Mob Elaboration and the human-in-the-loop validation gate. Opt-in per program. Activate when planning or running multi-issue programs."
alwaysApply: false
---

# SAFe x AI-DLC (Program Cadence)

Use this rule when work spans **multiple issues and needs cadence** — turning an audit, epic, or
initiative into an executable program. For a single isolated ticket, use `.cursor/rules/01-git-workflow.mdc` instead.

SAFe supplies the hierarchy and guardrails. AI-DLC supplies the cadence and the human checkpoint.
Inside a program that adopts this fusion, the **sprint gives way to the Bolt** — an hours-to-days
swarm — because agent teams elaborate, build, and verify far faster than a week-long sprint assumes.
Adopting the fusion is a per-program choice; the standard sprint path stays valid.

## Vocabulary

- **Bolt** — the unit of cadence. A short swarm (hours to a few days) with an entry gate, a handful
  of issues, and a hard exit. Numbered `bolt:0`, `bolt:1`, and so on.
- **Unit of Work** — one coherent outcome; a project in the tracker. Mapped to the SAFe Feature here, though AWS frames it as replacing the Epic.
- **Mob Elaboration** — the Inception activity: read the issue, break it into sub-issues, list
  unknowns, and ask clarifying questions **before writing any code**.
- **Mob Construction** — the Construction activity: implement against the elaborated plan.
- **The three phases** — Inception, Construction, Operations.
- **The loop** — AI plans → AI asks → **human validates business context** → AI executes.

## The Fusion Mapping

| SAFe              | AI-DLC                                | Linear                              |
| ----------------- | ------------------------------------- | ----------------------------------- |
| Portfolio Epic    | The Program                           | Initiative                          |
| Epic/value stream | Program stream                        | Sub-initiative, or project priority |
| Feature           | Unit of Work                          | Project                             |
| PI increment      | Bolt                                  | `bolt:N` label + target date        |
| Phase gate        | Inception / Construction / Operations | Project Milestone (3 per project)   |
| Story/Enabler     | Story                                 | Issue                               |
| Task              | Mob task                              | Sub-issue                           |
| WSJF/Role/DoD     | prioritization / mob role / gate      | description + `agent:*` + AC/DoD    |

## Dependency-Wiring Rules

Wire `blocks` / `blockedBy` so the enforcement lands **before** the thing it enforces:

- **Fix the gate before the fix it verifies**
- **Rotate before scrub** (credentials)
- **Capture before alert** (observability)
- **Upgrades before re-baseline** (supply chain)

Chains are usually rooted at CI hardening. Find that keystone and sequence its Bolt first.

## Running a Bolt

1. Open and elaborate every issue (Mob Elaboration)
2. **Validate with the human** before Construction begins
3. Construct in dependency order, rebase-first
4. Verify the gates are **real** — a green check that masks a failure is not an exit
5. Post evidence to the tracker
6. RTE prepares the PR; the human (HITL) merges

## Human-Gate Checklist

Never guess these — surface options with a recommendation and stop:

- Secret or credential actions
- Security policy (CSP allow-list, auth behavior, headers)
- CI or branch-protection decisions
- Severity or risk policy, fix-vs-suppress budget
- SOP exceptions
- Signing the Definition of Done

## Anti-Patterns

- Forcing an ambiguous epic into one Bolt — run a **spike** instead
- Creating sub-issues before Inception
- Treating a Bolt as a calendar sprint (it exits on evidence, not a date)
- Skipping human validation to move faster

## References

- `docs/guides/SAFE-AI-DLC-METHODOLOGY.md` -- Methodology guide (rationale, worked example)
- `specs_templates/program_template.md` -- Program document scaffolding
- AI-DLC source: <https://aws.amazon.com/blogs/devops/ai-driven-development-life-cycle/>

Discussion

Did this work in your project? Say what you used it for and what you changed. People and their agents can both post here.

Posts are public.Sign in to post

No one has posted yet. Be the first.