agentleFS
drivestream-labdrivestream-lab/prayog-skills/skills/development/initiative-feasibility/SKILL.md

initiative-feasibility

Skill2 starsChanged 4 months agoMITView on GitHub

After the dev team has written their spec slice from the PRD, review it against the current codebase — baseline, gaps, impact, test harness, architecture governance (ADR + MDC), risks, and 4-lane triage. Writes the report locally and emits Forge readiness for the open Draft spec PR while Gate 2 remains spec-pending. Use when the dev has drafted the spec and wants to check buildability before technical review and planning.

The file

Show its rules and text

We split this skill into the rules it gives an agent

15 rules. About 0 are common, 9 are distinctive, and 6 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Never skip a check in references/checks.md.

    line 28
  2. Distinctive:

    Don't fix — flag. Do not edit product source or product specs unless the user explicitly asks after the report.

    line 30
  3. Distinctive:

    Run T0–T5 control loop (Gather → Understand → Analyze → Plan → Execute → Verify).

    line 32
  4. Distinctive:

    Verify light source freshness before F1: product-spec header citations (H1 PRD digest, H2 scope digest, H3 map revision) and tip continuity.

    line 33
  5. Distinctive:

    Do not require mid-lane report digests as staleness SSOT.

    line 37
  6. Distinctive:

    Every open item must record lane, blocking, owner, status, required-by stage, default-if-deferred, evidence, and resolution reference.

    line 38
  7. Very specific:

    If a codegraph provider is available (MCP tool matching prayog-skills/references/codegraph-tool-contract.md, or a local CLI), prefer it for …

    line 47
  8. Distinctive:

    After dev updates spec — re-run on changed sections

    line 78
  9. Distinctive:

    If the repo was removed, deferred, or its scope digest changed, stop and emit the map's ripple action.

    line 85
  10. Very specific:

    Do not route PM/domain blockers to technical review via findings.

    line 112
  11. Distinctive:

    Never carry a prior feasibility finding forward across changed H1–H3 authority without explicitly re-evaluating that finding.

    line 124
  12. Very specific:

    Do not route engineering decisions to PM.

    line 143
  13. Very specific:

    When the invocation binds handoff_path (orchestrator / AgentRunner baton), also overwrite that path with the same handoff: envelope before …

    line 149
  14. Very specific:

    Follow prayog-skills/references/forge-side-effects.md#content-producers when this stage's pin has forge.commit_workspace other than …

    line 151
  15. Very specific:

    Put lane counts, NEW-ADR, selected outcome rationale, and ripple action in signals.

    line 158

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Initiative feasibility

Assess whether the dev team's spec slice for this initiative is buildable in this repo as it exists today. Do not implement — flag only.

The spec slice is written by the dev team after reading the PRD. This skill reviews it for gaps, codebase alignment, and architecture governance.

Pattern borrowed from awesome-copilot create-github-issues-for-unmet-specification-requirements (extract → search code → classify). Evaluation discipline aligned with validate-requirements.

NON-NEGOTIABLE

  1. Never skip a check in references/checks.md. Mark SKIPPED with reason.
  2. Evidence for every finding — spec quote + repo path (file, symbol, or test name).
  3. Don't fix — flag. Do not edit product source or product specs unless the user explicitly asks after the report.
  4. Dual output: chat summary + saved report file + 4-lane triage.
  5. Run T0–T5 control loop (Gather → Understand → Analyze → Plan → Execute → Verify).
  6. Verify light source freshness before F1: product-spec header citations (H1 PRD digest, H2 scope digest, H3 map revision) and tip continuity; G1 meta head + approval while Gate 1 still applies. See prayog-skills/references/artifact-write-contract.md. Stop on authority drift. Do not require mid-lane report digests as staleness SSOT.
  7. Every open item must record lane, blocking, owner, status, required-by stage, default-if-deferred, evidence, and resolution reference.
  8. Read-only content skill. Persist the feasibility report locally and fill handoff.forge for /commit-workspace (or Gateflow ForgeClient) onto the same Draft spec PR branch opened by /spec-draft. Do not create branches, commit, push, open PRs, apply labels, create issues, merge, run probes, edit product source, or clean up environments. Gate 2 label stays spec-pending — do not set spec-lgtm during feasibility.
  9. Do not implement product code. Spec-only artifacts; publication is Forge.
  10. If a codegraph provider is available (MCP tool matching prayog-skills/references/codegraph-tool-contract.md, or a local CLI), prefer it for architecture/impact/coverage questions. Always fall back to direct source_roots reads when unavailable — never block or change outcome selection on its absence.

Inputs

Gather before starting. Resolve paths from .harness/profile.yaml or references/layout-defaults.md.

Dual workspace (orchestrated / Gateflow): workspace = app coding root (spec slice + feasibility report). meta_workspace when bound = meta checkout for PRD / impact-map freshness. Do not invent a meta path when empty.

  1. Initiative spec slice — primary doc (REQUIRED); written by dev team, lives under product_spec_dir in workspace
  2. PRD — upstream requirements from <client>-meta/prd/ under meta_workspace when bound (OPTIONAL; for conformance context; link from spec header)
  3. Canonical impact map + approval evidence — (REQUIRED) exact revision and tech-lead review referenced by the spec header (resolve under meta_workspace when bound)
  4. As-built — implementation-status.md (REQUIRED)
  5. Tests — tests_readme, unit_tests_dir, live_verify_dir, and toolchain config from profile (REQUIRED)
  6. Source — modules under source_roots from profile (REQUIRED)
  7. Prior feasibility report — for incremental re-run (OPTIONAL)
  8. Layout — .harness/profile.yaml or references/layout-defaults.md
  9. rules_glob — workspace MDC rules (REQUIRED). Read before T2 Analyze.
  10. adr_dir — architecture decision records (REQUIRED). Run relevant-ADR pass per references/governance.md before T2 Analyze.
  11. Codegraph provider — OPTIONAL — see prayog-skills/references/codegraph-tool-contract.md

When to use

  • Dev has written spec slice and wants to check buildability before spec PR merge
  • After dev updates spec — re-run on changed sections
  • User asks: feasibility, impact, gap analysis, spec vs codebase

Process

  1. T0 Gather and freshness gate — inventory inputs; compare product-spec H1–H3 citations (and G1 when applicable) with live PRD/map/review. If the repo was removed, deferred, or its scope digest changed, stop and emit the map's ripple action. Feasibility report digests are walk-time only.
  2. T1 Understand — initiative id, spec branch, review objective
  3. T2 Analyze — read spec waves/capabilities; scan repo evidence; cross-reference rules_glob and relevant ADRs; flag spec wording that conflicts with MDC patterns or Accepted ADRs
  4. T3 Plan — which checks run (full vs incremental)
  5. T4 Execute — run checks F1–F14 per references/checks.md
  6. T5 Verify — save report locally; publish summary + 4-lane triage; select workflow outcome from the lane-to-outcome rubric; fill Forge readiness

Outcome selection (lane → workflow)

Map evidence to exactly one outcome declared for initiative-feasibility in pinned workflow.yaml. Prefer the first matching row:

OutcomeWhenNext (from workflow)
staleProduct-spec H1–H3 / G1 / tip authority driftspec-draft
failedExecution/analysis failure on otherwise valid inputsworkflow-stop
blockedExplicit approval/policy gate prevents progress (repo held, gate closed)spec-human-decision
needs-inputUnresolved blocking PM or domain item (missing product/SME answer)spec-human-decision
findingsUnresolved blocking PE / ADR / engineering item (Critical or Should-fix PE-lane)spec-technical-review
passZero unresolved blocking findings; informational/Verify/Gap observations may remainspec-technical-review

Both pass and findings enter /spec-technical-review before plan (pin SSOT). pass still means “no blocking PE findings at feasibility”; TDD may record N/A / light confirmation. Do not skip technical review on clean feasibility.

Do not route PM/domain blockers to technical review via findings. Do not emit pass while unresolved PE/ADR blockers remain. Informational observations are report signals — they do not alone select findings.

Impact-map revision handling

  • continue — scope digest unchanged; record the newer revision check.
  • re-draft / re-feasibility — scope digest changed; update the spec before running incremental feasibility.
  • hold / close — repo removed, deferred, blocked, or approval stale.
  • re-plan — dependency/build order changed; invalidate any existing plan.

Never carry a prior feasibility finding forward across changed H1–H3 authority without explicitly re-evaluating that finding. Feasibility report files are PURGE at initiative closure.

Output

Save to {reports_dir}/{feasibility_prefix}-{initiative}.md (from profile).

Use references/output-template.md.

Open items — 4-lane triage

End with four lanes (see output template):

  • PM questions — product scope, UX, priority; comment on meta PRD PR (plain English)
  • PE questions — engineering decisions, ADR gaps, test policy; comment on spec PR; resolved by /spec-technical-review
  • Domain clarifications — business source-of-truth; route to named SME (meta PRD PR or issue)
  • Auto-fixable — naming drift, inferred cross-references; record as findings/signals for later agent fix — do not mutate product source inside this skill

Do not route engineering decisions to PM. Apply the routing rubric in the spec-technical-review skill's references/governance.md.

Workflow handoff

  1. Append/emit the envelope from prayog-skills/references/handoff-envelope.md to the saved report. Use stage initiative-feasibility.
  2. When the invocation binds handoff_path (orchestrator / AgentRunner baton), also overwrite that path with the same handoff: envelope before exit. Leaving the baton empty is a failed stage for automated consumers. artifact.path remains the workspace skill output, not the baton path. See prayog-skills/references/handoff-envelope.md (Orchestrator baton).
  3. Derive next_candidates and human_checkpoint from pinned root workflow.yaml for (stage: initiative-feasibility, outcome) per prayog-skills/references/handoff-envelope.md (Derive from pinned workflow). Set human_checkpoint: true only when the resolved next node's type is human-checkpoint — never because the artifact "should be reviewed."
  4. Follow prayog-skills/references/forge-side-effects.md#content-producers when this stage's pin has forge.commit_workspace other than disabled or next is an external-action with forge.requires — fill handoff.forge / recommend /commit-workspace; do not treat local CLI as skill success. Content skill success ≠ report published.

Transitions: pinned root workflow.yaml for this stage (SSOT) — including pass vs findings routes. Human or agent may run this skill; legality and auto-dispatch follow dispatch + delivery contract + latest handoff.

Put lane counts, NEW-ADR, selected outcome rationale, and ripple action in signals; finding ids belong in blockers as FF-* (never bare F-12). next_candidates never authorize invoke.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.