fleet-program

name	fleet-program
description	Orchestrate autonomous-fleet missions on one repo — sequential chains and conditional campaign DAGs with if-outcome edges. Reads fleet-outcome YAML from each mission's readiness doc to branch (e.g. audit then tests if no P0s, else dependency-update). One mission active at a time per repo; cross-repo parallel via separate sessions. Does not run parallel missions on the same repo. Use for "repo health program", "audit then test", "docs then bugs if needed", mission chains, fleet campaign, ship with proof, align then ship, quality gate. Install from github.com/ravidsrk/autonomous-fleet. Trigger on: "fleet program", "fleet campaign", "mission chain", "if P0 then", "repo health", "conditional fleet run", "ship safely", "ship with proof", "finish stalled product", "align then ship", "production ready", "quality gate".
license	MIT
compatibility	Requires git and gh CLI; install mission skills via npx skills
metadata	{"author":"ravidsrk","version":"1.2.4","fleet-component":"program"}

fleet-program

Meta-skill for mission-level orchestration: linear programs and conditional campaign DAGs. You are the PROGRAM COORDINATOR — the engine, one level up for missions. Same discipline: file ledger, frozen outcomes, no-stop autonomy, one active mission per repo.

Required skills

1. autonomous-fleet-core — references/engine.md, references/composition.md, references/fleet-outcome.md, references/runtime-goals.md
2. One runtime adapter: autonomous-fleet-adapter-orca, autonomous-fleet-adapter-claude-code, autonomous-fleet-adapter-grok, or autonomous-fleet-adapter-codex

Install missions via npx skills add before starting.

Load only the active mission's skill while that mission runs — never two mission skills at once.

Optional skills

Skill	Activate when	If unavailable
autonomous-fleet	Vague intent; need catalog	Pick preset from references/programs.md or references/campaigns.md

Program ledger

<LEDGER_DIR>/fleet-program-progress.md (LEDGER_DIR defaults to docs/; docs-site repos relocate it — engine SELF-ORIENTATION step 6):

# Fleet program progress

MODE: <linear | campaign | parallel_repos>
CAMPAIGN: <id>
PHASE: <PLANNING | NODE-<id> | DONE | BLOCKED>
ACTIVE_MISSION: <mission-id | none>
CURRENT_NODE: <node-id | none>
BASE: <branch>

## Campaign spec
(paste YAML from campaigns.md or user)

## Last fleet-outcome
(paste parsed summary from last readiness doc)

## Node status
| Node | Mission | Status | Readiness doc |
|------|---------|--------|---------------|

## Runtime goal

SCOPE: campaign
CONDITION: |
  Campaign <id> DONE: <LEDGER_DIR>/fleet-program-progress.md PHASE is DONE,
  every node DONE or SKIPPED, each readiness doc has valid fleet-outcome YAML.
HOST: <adapter runtime>
SET_AT: <timestamp>
LAST_UPDATE: <progress>

## Handoff notes

Status per node: PENDING | RUNNING | DONE | SKIPPED.

Choose mode

Mode	When	Spec
linear	Ordered list, no branches	programs.md table → implicit always edges
campaign	if branches on outcomes	campaigns.md DAG YAML
parallel_repos	Same mission on different repos	campaigns.md parallel_repos — separate sessions, aggregate at end

Default vague intent → repo-health campaign (linear DAG). Security / pre-merge → ship-with-proof or secure-ship. Stalled product (explicit Tier 3) → align-then-ship. Acceptance / readiness → quality-gate. Community hooks: community-skills.md.

Planning

1. SELF-ORIENT (core engine).
2. Parse user request → linear queue OR campaign YAML OR parallel_repos.
3. Write spec + ledger; record in DECISIONS.md.
4. BASE: <BRANCH_PREFIX><campaign-id>-base off default branch at HEAD (first node).
5. SET_GOAL (campaign scope) per runtime-goals.md — condition must reference <LEDGER_DIR>/fleet-program-progress.md and readiness validation.

Runtime goal binding

Step	Action
Campaign start	SET_GOAL(campaign_done_condition); write ## Runtime goal in program ledger
Each node start	SET_GOAL(mission_done_condition) — replaces campaign goal for this session
Node complete	UPDATE_GOAL("node <id> done: <fleet-outcome summary>"); run python3 <SUBSTRATE>/validate_fleet_outcome.py on readiness doc (per SUBSTRATE RESOLUTION; skip when SUBSTRATE=none)
Campaign complete	GOAL_COMPLETE only when PHASE: DONE in file + all validations pass
Blocked node	GOAL_BLOCKED if fleet-outcome.status == blocked and no retry

Mission goal template (substitute mission id, ledger, readiness, metrics):

Mission <mission-id> DONE: <LEDGER_DIR>/<mission>-progress.md all task flags true,
<LEDGER_DIR>/<mission>-readiness.md with fleet-outcome.status done and mission metrics satisfied,
the readiness fleet-outcome validates (python3 <SUBSTRATE>/validate_fleet_outcome.py per the engine's SUBSTRATE RESOLUTION; skip recorded in the readiness doc when SUBSTRATE=none), all PRs merged into BASE.

Unattended CI (framework clone only): ./scripts/run-campaign.sh <grok|claude|codex> --preset repo-health --max-turns N (or --campaign docs/<campaign>.yaml; add --dry-run to plan only)

Per-mission loop (single repo)

1. Set CURRENT_NODE, ACTIVE_MISSION, PHASE: NODE-<id>.
2. SET_GOAL for active mission (see template above).
3. Activate only that mission skill.
4. Run mission to DONE (mission ledger + readiness with fleet-outcome YAML).
5. Parse outcome: read YAML frontmatter from readiness doc per fleet-outcome.md. Store in Last fleet-outcome. If missing, extract metrics from readiness prose and log a warning in DECISIONS.md.
6. UPDATE_GOAL with node summary; mark node DONE; update handoff (deferrals → next mission discovery tasks).
7. Pick next node:
   • Linear: next row in queue.
   • Campaign: from edges[current_node], evaluate each if in order; first true edge wins. Expressions use fleet-outcome.metrics.*, top-level fields, and always. See campaigns.md.
   • No matching edge → PHASE: DONE.
8. If fleet-outcome.status == blocked → PHASE: BLOCKED; GOAL_BLOCKED; stop chain unless mission rules allow retry.
9. When no next node → set PHASE: DONE; GOAL_COMPLETE after file validation.

Conditional expression evaluator

if value	True when
always	Always
p0_open > 0	metrics.p0_open compares (same for any metric key)
p0_open == 0	Equality
code_bug_findings > 0	From doc-sync metrics
status == blocked	Top-level status
deferred_missions contains bug-batch	Any deferral id matches

Unknown expression → log in DECISIONS.md, skip edge (do not guess).

Parallelism

Case	Rule
Same repo, two missions	Forbidden — no shared lock manager
Tasks inside active mission	Mission hot-file + placement rules
Different repos	parallel_repos: one coordinator loop per repo OR user runs separate sessions

Autonomy (program level)

Same as core: do not stop between missions; do not ask "continue to next mission?"; circuit-breaker on a node → SKIPPED + DECISIONS.md, then evaluate if campaign can proceed.

DONE

All nodes DONE or SKIPPED, or PHASE: DONE / BLOCKED with reason. FINAL report: campaign spec, per-node outcomes (fleet-outcome summaries), readiness links, combined deferrals.

Safe defaults

• First-time repo: repo-health campaign.
• Security intent: secure-ship or audit-branch (conditional).
• Tier 3 missions only when user explicitly requests.