portfolio-pdlc

Turn any value stream — a PDLC, an SDLC, a content pipeline, an AI use-case pipeline, an operational request stream — into a flow-driven agentic loop that runs on plain files. Use when choosing where to apply agentic workflow, defining or ingesting a definition of workflow (steps, graph, exit evidence, delegate model per step, run model per step, measures), scaffolding it as markdown, orchestrating it agentically one move per cycle, integrating it with GitHub or a Kanban board, or inspecting and adapting the loop itself. Entry point and operating loop for the flow-driven-* skill family; for portfolio-level investment management use portfolio-pdlc instead.

Identify the value streams an organisation actually runs — operational (a request stream with someone waiting) and development (building something that did not exist) — and choose which one to make agentic first, with a scored, reversible recommendation. Use before defining any workflow, when an org asks "where should we apply AI/agents", when an agentic programme has produced demos but no change in outcomes, or when picking the next stream after the first one is running. Part of the flow-driven family.

Define the workflow for a value stream by discovering it, deriving it, or adapting it — interview people about how work actually happens today and match it against archetypes; or, when no workflow exists, derive one from first principles and explain why working this way is worth it. Produces the unit of value, the steps and graph (branches, back-edges, sub-graphs), exit evidence per step, pull policies and WIP limits, classes of service, decision points, delegate and run model per step, and cadences. Use when standing up a new agentic workflow, surfacing an implicit one, or reworking a definition that no longer matches reality. Part of the flow-driven family.

Run the meta-loop — inspect and adapt the workflow itself. Probe the flow (constraint drift, rework hotspots, WIP theatre, decision starvation), the evidence (watermelon steps, checks that always pass, self-graded work), the delegation (rung mismatches, escalation droughts, context rot, platform stall), and the definition of workflow (policy contradictions, silent drift, measures nobody steers with) — then capture improvement bets with benefit hypotheses and kill criteria instead of changing anything on the spot. Use on the meta-loop cadence, after recurring findings, or when the board is clean enough to earn self-improvement. Loop 2 of the flow-driven family.

Adopt a workflow that already exists — GitHub Spec Kit, Kiro, a bespoke prompt/script/cron harness, Claude Code or Codex skill libraries, a Jira/Linear tracker, or an n8n/Temporal automation — and map it into a definition of workflow without rewriting what works. Use when an organisation already has agentic or process machinery and wants it made flow-driven, when documented and actual process have drifted apart, or when deciding whether to wrap, extend, or retire an existing harness. Produces the mapped workflow plus a delta report. Part of the flow-driven family.

Decide what a workflow should measure and wire it in — flow measures (WIP, age, cycle time, throughput, rework), evidence measures (coverage, first-pass yield, escape rate, check agreement), leverage measures (delegate mix, human minutes per item, escalation rate, cost per item), and the outcome measure the stream exists for — with each measure named on the steps that feed it and a baseline taken before anything changes. Use when defining a workflow, when a board has numbers nobody steers with, or before making any workflow-improvement bet. Part of the flow-driven family.

Bind a flow workspace to wherever the work is actually managed — GitHub Issues and Projects, Jira, Linear, an internal Kanban dashboard, a spreadsheet, or chat — by choosing a system-of-record mode (files-only, mirror, adopt, split), writing a field-ownership map with exactly one writer per field, and using the board export plus the harness's own tools to sync. Use when leadership needs the flow where they already look, when a tracker already holds the work, or when a sync has started fighting the workspace. Part of the flow-driven family.

Advance a flow workspace agentically — validate the definition, regenerate the board, take the single highest-leverage move from the deterministic run card (overdue decision, broken contract, blocked item, missing exit evidence, ready-to-pull, over-WIP, aging, or intake), run that step under its contract, verify the exit evidence with something independent, and record the transition. Use to run a cycle, work the flow on a cadence, drive an item to its next decision, or operate the loop headlessly. Loop 1 of the flow-driven family.

Materialise a definition of workflow as a working flow workspace — workflow.md with its graph, one step contract per step with real stubs, item cards, measures, integration bindings, platform context-pack and check stubs, and the first generated board — so a human or agent can take over from a running skeleton rather than a blank page. Use after defining or ingesting a workflow, when standing up a second stream, or when an existing workspace needs missing pieces filled in. Part of the flow-driven family.

Help someone choose how to frame the intent behind an initiative — Lean Product Canvas, a SAFe epic hypothesis statement, a feature template, or something they shape themselves — and then configure a local coaching skill around the one they picked. Use when an initiative is fuzzy and someone asks "what template should we use", when a team has inherited a framing artifact that does not fit the work, when an AI use case or business initiative needs an outcome frame rather than a product one, or before running sniff-test on an initiative that has no framing artifact at all. Teaches the trade-offs between the options; does not reproduce any of them.

White-box assessment of a spec-driven harness or coding-agent configuration from a public GitHub repository, local directory, ZIP archive, or effective instruction hierarchy. Use when a CTO, VPE, or AI adoption leader wants evidence-backed feedback on outcome discovery, trustworthy delivery, in-use validation, and how the harness improves its own loop. For a produced spec as black-box evidence, use pdlc-readiness instead.

Assess a produced spec or spec bundle as black-box evidence of outcome orientation and PDLC readiness. Use when a CTO, VPE, or AI adoption leader wants to improve a spec-driven harness from its outputs, separating current-spec repairs from reusable workflow changes. For white-box inspection of a harness repository or agent configuration, use pdlc-harness-readiness.

Run a portfolio-level, product-oriented PDLC (Explore → Discovery → Plan/Commit → Execute → Rollout → BAU, with humans keeping the investment decisions) as an operating loop that leaves the portfolio sharper every cycle, over a plain-file workspace. Use when standing up, operating, or resuming a portfolio of initiatives/epics — seeing where everything is, advancing initiatives with evidence, improving artifact quality and outcome orientation, and continuously improving the process and portfolio construction themselves. Entry point and operating loop for the portfolio-pdlc-* skill family; for one-off diagnostics of a single initiative use sniff-test instead.

Advance ONE initiative (or improvement bet) through the portfolio PDLC — do the stage-appropriate work that grows confidence, and prepare (never make) the human decision at the next boundary. Use when an item is aging, a discovery timebox lapsed without a recommendation, a decision is due or overdue, or a card is ready to move stage. The loop-1 skill of the portfolio-pdlc family.

The sensing pass of the portfolio-pdlc operating system. Use when asking "where is everything really, and is it outcome-oriented or activity-oriented?" — full portfolio reads, review/investment-council prep, stale-status sweeps, or restoring board integrity after contract violations. Verifies stages against evidence, classifies every card on the input→impact ladder, reads flow and risk balance, and outputs a review brief with ranked candidate moves. Composes the sniff-test and sniff-test-portfolio skills included in this repo.

Probe the portfolio operating system itself — both the PDLC process (flow boundaries, altitude, WIP, discovery discipline, measure-and-learn) and the portfolio construction/topology (dependency concentration, platform extraction, team-around-outcome options) — and capture improvement ideas as spec-driven bets in the improvements/ backlog instead of implementing them on the spot. Use when the board is healthy enough to earn self-improvement, after recurring findings, or on an improvement cadence. Loops 3+4 of the portfolio-pdlc family; focus:process, focus:topology, or focus:auto.

Run a deterministic Monte Carlo simulation of flow through the portfolio PDLC — baseline from the workspace's own flow history, plus what-if scenarios (WIP limit changes, arrival-rate shaping, dependency-tax reduction from platform extraction or team reorganization). Use as the Discovery move for improvement bets before anyone changes policies or reorganizes teams, or to set Service Level Expectations. Advanced member of the portfolio-pdlc family; produces confidence-labeled forecasts, never point predictions.

Improve the quality, outcome orientation, and derisking of ONE initiative's artifacts in the portfolio PDLC — rewrite activity-framed cards into outcome hypotheses with leading indicators, sharpen riskiest assumptions with evidence/opinion tags, fix derisking-approach mismatches, and capture at most one durable learning. Use on cards flagged activity-oriented, thin on evidence, missing indicators, or carrying committed money with open risks. The loop-2 skill of the portfolio-pdlc family.

Stand up a portfolio-pdlc workspace from whatever an organization already has — folders of decks and docs, a Jira/Align export, a spreadsheet of initiatives, or a blank page. Use when wiring the portfolio PDLC for the first time, onboarding an engagement's existing initiative material into the operating system, or re-wiring after a major re-scoping. Produces the charter, workflow definition, initiative cards, and first generated board. Part of the portfolio-pdlc family.

Run a lightweight, evidence-based "sniff test" on a single initiative (epic, project, or Tier-1 bet) to find where its clarity is insufficient for its stage in the Portfolio PDLC lifecycle, so you can drive the right investment/commit and derisking conversations. Best run pointed at where the initiative is managed (its folder/workspace, Confluence space, or Jira epic) or given its Lean Product Canvas plus a recent sponsor update; with little input it asks for context first. Use when pressure-testing an initiative before a Plan/Commit decision or Execute/Build commitment, checking whether a Lean Product Canvas and Discovery plan match the real risks, or coaching on leading indicators and risk-appropriate derisking. For multi-initiative board patterns, use the sniff-test-portfolio add-on.