SAFe x AI-DLC Skill
📋 TEMPLATE: This skill uses {{TICKET_PREFIX}} and {{MAIN_BRANCH}} as placeholders.
Replace with your project's ticket prefix (e.g., WOR, PROJ) and base branch (e.g., dev, main).
Purpose
Encode the SAFe x AI-DLC fusion so agents structure multi-issue programs consistently: SAFe supplies
the hierarchy and guardrails, AI-DLC supplies the cadence and the human checkpoint. Inside a program
that adopts this fusion, sprints give way to Bolts, because agent teams elaborate, build, and
verify in hours rather than weeks. Adopting the fusion is a per-program choice; the standard sprint
path stays valid, and single isolated tickets route to safe-workflow instead.
When This Skill Applies
Invoke this skill when:
- Turning an audit, epic, or initiative into an executable program
- Structuring work in Linear as initiative, projects, milestones, issues, and sub-issues
- Breaking an initiative into Units of Work and sequencing them into Bolts
- Wiring dependencies so gates land before the fixes they verify
- Running a Bolt: elaborate, validate with the human, construct, verify gates are real, evidence, merge
Do NOT use it for a single isolated ticket with no program context — use safe-workflow for that.
The Fusion Model
SAFe gives hierarchy, WSJF, roles, DoD, and dependencies. AI-DLC gives cadence (Bolts) and the
human-in-the-loop. See docs/guides/SAFE-AI-DLC-METHODOLOGY.md for the full rationale.
| 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:* labels + AC/DoD |
The loop: AI plans → AI asks clarifying questions → HUMAN validates business context → AI
executes. Never skip the human validation step.
How to Structure Linear
Build top-down:
- Initiative = the program. One per program.
- Projects = Units of Work. One coherent outcome each.
- Milestones = the three AI-DLC phases (Inception, Construction, Operations). Create all three
per project.
- Issues = Stories. Each carries the issue template below. Label with
bolt:N and agent:*
(lead role).
- Sub-issues = Mob Elaboration tasks. Create during Inception, not before.
- Dependencies =
blocks / blockedBy edges forming the dependency DAG.
Program streams map to sub-initiatives where the Linear plan supports them; where it does not,
encode the stream as project priority instead (highest-risk stream = Urgent/High).
Use linear-sop for the exact MCP tool calls, label prerequisites, and evidence templates.
Issue Template
Every issue description block contains:
Header: <ID> — <title> [{{TICKET_PREFIX}}-XXX]
WSJF: <score> MoSCoW: <Must|Should|Could|Won't>
Finding: <source refs, e.g. audit finding IDs>
Phase: <Inception | Construction | Operations>
Bolt: <bolt:N>
Role: <agent:* lead>
AC: - acceptance criterion (testable)
DoD: - merged to {{MAIN_BRANCH}}; relevant gate REAL and green; evidence in Linear;
no silent suppression
Deps: blocks: [...] blockedBy: [...]
Sub-issues: (Mob Elaboration tasks, added during Inception)
Dependency-Wiring Rules
Wire blocks / blockedBy so the enforcement lands before the thing it enforces:
- Fix the gate before the fix it verifies. Harden CI or un-mask a check before landing the fix
that check is supposed to catch.
- Rotate before scrub. Rotate a leaked credential before scrubbing it from history.
- Capture before alert. Error capture or health check blocks the alerter and the poller built on
top of it.
- Upgrades before re-baseline. Framework and SDK upgrades block audit re-baselining, which blocks
the scheduled recurring audit.
Dependency chains are usually rooted at CI hardening — one keystone issue tends to unblock a whole
stream. Identify it and sequence its Bolt first.
Running a Bolt
- Open and elaborate every issue in the Bolt (Mob Elaboration: sub-issues, unknowns, questions).
- Validate with the human on all business, security, and policy decisions before Construction.
- Construct in dependency order on
{{TICKET_PREFIX}}- branches with SAFe commits, rebase-first.
- Verify gates are real — a green check that still masks a failure is not an exit.
- Post evidence to Linear (dev/staging/done) per
linear-sop.
- Merge — RTE prepares the PR; the human (HITL) merges. Exit = merged + real gates green +
evidence posted.
Human-Gate Checklist
Route to the human, with options and a recommendation, whenever a decision is:
- A secret or credential action (what to rotate, when, blast radius)
- A security policy: CSP allow-list, auth behavior, headers
- A CI or branch-protection decision: which checks become required
- A severity or risk policy: dependency-audit thresholds, fix-vs-suppress budget
- An SOP exception (e.g. a migration deviating from the documented process)
- Signing the Definition of Done
If a decision changes business, security, or risk posture, it is the human's — surface, recommend,
stop.
Anti-Patterns (Do NOT use)
Forcing an ambiguous epic into one Bolt (run a spike instead — elaboration is the hard part)
Creating sub-issues before Inception (Mob Elaboration produces them, not the other way round)
Bolts as calendar sprints (a Bolt exits on evidence, not on a date)
Skipping human validation to move faster (the loop is the method; without it this is just SAFe)
Treating a green check as an exit (verify the gate enforces, not that it merely passed)
Authoritative References
Routes To
linear-sop — MCP tool calls, program structure, issue creation, evidence templates
safe-workflow — branch, commit, and PR conventions; HITL merge
orchestration-patterns — multi-agent swarm coordination for a Bolt
pattern-discovery — find existing code and doc patterns before constructing
1---2name: safe-ai-dlc3description: Plan and run programs using the SAFe x AI-DLC fusion. Use when turning an audit, epic, or initiative into Linear structure (initiative, projects, milestones, issues, sub-issues), organizing work as Units of Work and Bolts, wiring a dependency DAG, or running a Bolt swarm with a human-in-the-loop gate. Applies when work spans multiple issues and needs cadence, not just a single ticket.4user-invocable: false5allowed-tools: Read, Grep, Glob6metadata:7 source: https://skillmd.com/skills/bybren-llc/safe-ai-dlc8---910# SAFe x AI-DLC Skill1112> **📋 TEMPLATE**: This skill uses `{{TICKET_PREFIX}}` and `{{MAIN_BRANCH}}` as placeholders.13> Replace with your project's ticket prefix (e.g., `WOR`, `PROJ`) and base branch (e.g., `dev`, `main`).1415## Purpose1617Encode the SAFe x AI-DLC fusion so agents structure multi-issue programs consistently: SAFe supplies18the hierarchy and guardrails, AI-DLC supplies the cadence and the human checkpoint. Inside a program19that adopts this fusion, sprints give way to **Bolts**, because agent teams elaborate, build, and20verify in hours rather than weeks. Adopting the fusion is a per-program choice; the standard sprint21path stays valid, and single isolated tickets route to `safe-workflow` instead.2223## When This Skill Applies2425Invoke this skill when:2627- Turning an audit, epic, or initiative into an executable program28- Structuring work in Linear as initiative, projects, milestones, issues, and sub-issues29- Breaking an initiative into Units of Work and sequencing them into Bolts30- Wiring dependencies so gates land before the fixes they verify31- Running a Bolt: elaborate, validate with the human, construct, verify gates are real, evidence, merge3233Do NOT use it for a single isolated ticket with no program context — use `safe-workflow` for that.3435## The Fusion Model3637SAFe gives hierarchy, WSJF, roles, DoD, and dependencies. AI-DLC gives cadence (Bolts) and the38human-in-the-loop. See `docs/guides/SAFE-AI-DLC-METHODOLOGY.md` for the full rationale.3940| SAFe | AI-DLC | Linear |41| ------------------ | ------------------------------------- | ----------------------------------------- |42| Portfolio Epic | The Program | Initiative |43| Epic/value stream | Program stream | Sub-initiative, or project priority |44| Feature | Unit of Work | Project |45| PI increment | Bolt | `bolt:N` label + target date |46| Phase gate | Inception / Construction / Operations | Project Milestone (3 per project) |47| Story/Enabler | Story | Issue |48| Task | Mob task | Sub-issue |49| WSJF/Role/DoD | prioritization / mob role / gate | description + `agent:*` labels + AC/DoD |5051**The loop:** AI plans → AI asks clarifying questions → **HUMAN validates business context** → AI52executes. Never skip the human validation step.5354## How to Structure Linear5556Build top-down:57581. **Initiative** = the program. One per program.592. **Projects** = Units of Work. One coherent outcome each.603. **Milestones** = the three AI-DLC phases (Inception, Construction, Operations). Create all three61 per project.624. **Issues** = Stories. Each carries the issue template below. Label with `bolt:N` and `agent:*`63 (lead role).645. **Sub-issues** = Mob Elaboration tasks. Create during Inception, not before.656. **Dependencies** = `blocks` / `blockedBy` edges forming the dependency DAG.6667Program streams map to **sub-initiatives** where the Linear plan supports them; where it does not,68encode the stream as **project priority** instead (highest-risk stream = Urgent/High).6970Use `linear-sop` for the exact MCP tool calls, label prerequisites, and evidence templates.7172## Issue Template7374Every issue description block contains:7576```text77Header: <ID> — <title> [{{TICKET_PREFIX}}-XXX]78WSJF: <score> MoSCoW: <Must|Should|Could|Won't>79Finding: <source refs, e.g. audit finding IDs>80Phase: <Inception | Construction | Operations>81Bolt: <bolt:N>82Role: <agent:* lead>83AC: - acceptance criterion (testable)84DoD: - merged to {{MAIN_BRANCH}}; relevant gate REAL and green; evidence in Linear;85 no silent suppression86Deps: blocks: [...] blockedBy: [...]87Sub-issues: (Mob Elaboration tasks, added during Inception)88```8990## Dependency-Wiring Rules9192Wire `blocks` / `blockedBy` so the enforcement lands **before** the thing it enforces:9394- **Fix the gate before the fix it verifies.** Harden CI or un-mask a check before landing the fix95 that check is supposed to catch.96- **Rotate before scrub.** Rotate a leaked credential before scrubbing it from history.97- **Capture before alert.** Error capture or health check blocks the alerter and the poller built on98 top of it.99- **Upgrades before re-baseline.** Framework and SDK upgrades block audit re-baselining, which blocks100 the scheduled recurring audit.101102Dependency chains are usually rooted at CI hardening — one keystone issue tends to unblock a whole103stream. Identify it and sequence its Bolt first.104105## Running a Bolt1061071. **Open and elaborate** every issue in the Bolt (Mob Elaboration: sub-issues, unknowns, questions).1082. **Validate with the human** on all business, security, and policy decisions before Construction.1093. **Construct** in dependency order on `{{TICKET_PREFIX}}-` branches with SAFe commits, rebase-first.1104. **Verify gates are real** — a green check that still masks a failure is not an exit.1115. **Post evidence** to Linear (dev/staging/done) per `linear-sop`.1126. **Merge** — RTE prepares the PR; the human (HITL) merges. Exit = merged + real gates green +113 evidence posted.114115## Human-Gate Checklist116117Route to the human, with options and a recommendation, whenever a decision is:118119- A **secret or credential** action (what to rotate, when, blast radius)120- A **security policy**: CSP allow-list, auth behavior, headers121- A **CI or branch-protection** decision: which checks become required122- A **severity or risk policy**: dependency-audit thresholds, fix-vs-suppress budget123- An **SOP exception** (e.g. a migration deviating from the documented process)124- **Signing the Definition of Done**125126If a decision changes business, security, or risk posture, it is the human's — surface, recommend,127stop.128129## Anti-Patterns (Do NOT use)130131```text132Forcing an ambiguous epic into one Bolt (run a spike instead — elaboration is the hard part)133Creating sub-issues before Inception (Mob Elaboration produces them, not the other way round)134Bolts as calendar sprints (a Bolt exits on evidence, not on a date)135Skipping human validation to move faster (the loop is the method; without it this is just SAFe)136Treating a green check as an exit (verify the gate enforces, not that it merely passed)137```138139## Authoritative References140141- `docs/guides/SAFE-AI-DLC-METHODOLOGY.md` -- Methodology guide (rationale, worked example)142- `specs_templates/program_template.md` -- Program document scaffolding143- **AI-DLC source**: <https://aws.amazon.com/blogs/devops/ai-driven-development-life-cycle/>144145## Routes To146147- `linear-sop` — MCP tool calls, program structure, issue creation, evidence templates148- `safe-workflow` — branch, commit, and PR conventions; HITL merge149- `orchestration-patterns` — multi-agent swarm coordination for a Bolt150- `pattern-discovery` — find existing code and doc patterns before constructing