kushi-agents 3.4.1

package/plugin/prompts/consolidate.prompt.md

@@ -0,0 +1,21 @@
+---
+name: consolidate
+description: Merge all contributors' weekly streams into _Consolidated/ for a given window.
+---
+
+# /consolidate
+
+Route to `@Kushi consolidate <project> <window>`.
+
+Default window: `last 7 days`.
+
+Reads:
+- `<engagement-root>/<project>/Evidence/<each-alias>/**/stream/**` for the window
+- `<engagement-root>/<project>/Evidence/contributors.yml`
+
+Writes:
+- `<engagement-root>/<project>/Evidence/_Consolidated/<YYYY-MM-DD>_consolidated.md` (Monday of the week)
+
+Provenance tags retained (source: <alias>/<source>/<file>) per `citation-ledger.instructions.md`.
+
+Delegates to `consolidate-evidence` skill.

package/plugin/prompts/fde-intake.prompt.md

@@ -0,0 +1,41 @@
+---
+name: fde-intake
+description: Author or update the FDE Intake document — first artifact in an FDE engagement. Answers the 6 canonical FDE Intake questions, grounded in project Evidence + the FDE reference pack.
+---
+
+# /fde-intake
+
+Route to `@Kushi fde-intake <project>`.
+
+Authors (first time) or updates (subsequent runs) the **FDE Intake** at `<engagement-root>/<project>/Reports/00-FDE-Intake-<project>.md` — the first artifact in any FDE engagement. Read-only against Evidence; no outbound.
+
+## Syntax
+
+```
+@Kushi fde-intake <project>
+```
+
+## What it does
+
+- Loads project Evidence (read-only) + the FDE reference pack (`intake-questions.md`, `report-doctrine.md`, `core-fde-reference.md`).
+- Applies all 9 doctrine rules (source resolution gate, recency precedence, CRM-field-vs-confirmed-fact, validation warnings protocol).
+- Answers the 6 canonical FDE Intake questions:
+  1. Who is the customer + business problem?
+  2. What is being asked of FDE?
+  3. Why FDE (vs partner / product / advisory)?
+  4. When does this need to land + is the timeline realistic?
+  5. How is success measured (customer's own ROI)?
+  6. Who catches the outcome after FDE leaves?
+- Surfaces commercial summary (MACC / ECIF / funding / legal posture), MS + customer stakeholders, outcome KPIs, top risks (8 reference-pack categories), anti-patterns triggered.
+- Embeds inline `> ⚠️ VALIDATION WARNING — Rule X.Y` blockquotes wherever Evidence is silent.
+- Records per-source coverage in a `## Source Coverage` footer section.
+
+## Output
+
+`<engagement-root>/<project>/Reports/00-FDE-Intake-<project>.md`. Re-runs preserve user-authored prose; only Evidence-derived sections are overwritten. A `## Revision log` entry records each run.
+
+## See also
+
+- `skills/fde-intake/SKILL.md` for the full step-by-step.
+- `@Kushi fde-triage <project>` — produces the 7-file triage bundle that builds ON the intake.
+- `@Kushi fde-report <project> stage-readiness` — separate "should we advance the stage?" check.

package/plugin/prompts/fde-report.prompt.md

@@ -0,0 +1,46 @@
+---
+name: fde-report
+description: Generate an FDE-shaped engagement report in one of five shapes (weekly default / short / long / fitness / stage-readiness), grounded in project Evidence + the FDE reference pack. Read-only; no outbound.
+---
+
+# /fde-report
+
+Route to `@Kushi fde-report <project> [shape]`.
+
+Generates an FDE-shaped engagement report. Read-only against Evidence; no outbound.
+
+## Syntax
+
+```
+@Kushi fde-report <project>                       # weekly (default)
+@Kushi fde-report <project> weekly
+@Kushi fde-report <project> short                 # customer-facing, sanitized
+@Kushi fde-report <project> long                  # handoff / comprehensive
+@Kushi fde-report <project> fitness               # 10-row FDE fitness scorecard
+@Kushi fde-report <project> stage-readiness       # advance-the-stage check
+```
+
+## Shapes
+
+| Shape | Audience | Length |
+|---|---|---|
+| `weekly` | Internal MS leads | 1–2 pages |
+| `short` | Customer sponsor (sanitized — no MACC/staffing/CRM IDs) | 1 page |
+| `long` | Incoming FDE crew (handoff context dump) | 3–5 pages |
+| `fitness` | FDE leadership — 10-row fitness scorecard + verdict | 1 page |
+| `stage-readiness` | FDE leadership — should we advance the stage? | 1 page |
+
+## Notes
+
+- **`fitness`** shape ≡ bundle file 01 of `fde-triage` (use standalone when you want only the scorecard).
+- **`stage-readiness`** answers "should we advance to the next FDE stage?" — distinct from the triage bundle's **mobilization readiness** (file 04), which asks "can the crew start today?". Run both for a stage gate.
+
+## Output
+
+`<engagement-root>/<project>/Reports/fde-<shape>-<YYYY-MM-DD>.md`. Skill applies all 9 doctrine rules and embeds validation warnings inline.
+
+## See also
+
+- `skills/fde-report/SKILL.md` for the full step-by-step.
+- `@Kushi fde-intake <project>` — precursor first-artifact intake document.
+- `@Kushi fde-triage <project>` — full 7-file triage bundle (analysis + fitness + risk + 6Q + mobilization readiness + executive readout + reuse lens + validation warnings).

package/plugin/prompts/fde-triage.prompt.md

@@ -0,0 +1,46 @@
+---
+name: fde-triage
+description: Produce the full FDE Triage bundle — 7 companion files (analysis, fitness, risk, 6Q, mobilization readiness, executive readout, global reuse, validation warnings) at <project>/Reports/triage/<YYYY-MM-DD>/.
+---
+
+# /fde-triage
+
+Route to `@Kushi fde-triage <project>`.
+
+Produces the **FDE Triage bundle** — 7 files saved to `<engagement-root>/<project>/Reports/triage/<YYYY-MM-DD>/`. Read-only against Evidence; no outbound.
+
+## Syntax
+
+```
+@Kushi fde-triage <project>
+```
+
+## What's in the bundle
+
+| # | File | Purpose |
+|---|---|---|
+| 00 | `00-fde-analysis.md` | Concise FDE-oriented project analysis — bundle entry point |
+| 01 | `01-fde-fitness.md` | 10-row FDE fitness scorecard + verdict |
+| 02 | `02-risk-analysis.md` | Risks bucketed into 8 reference-pack categories |
+| 03 | `03-6Q.md` | 6-question engagement framing |
+| 04 | `04-readiness-checklist.md` | **Mobilization readiness** (distinct from stage readiness) |
+| 05 | `05-executive-consolidated-report.md` | Leadership-friendly combined readout |
+| 06 | `06-global-opportunity-and-reuse.md` | Repeatability + commercialization lens |
+| 07 | `07-validation-warnings-checklist.md` | Central tracker of all warnings (open / resolved / NA) |
+
+## Important — mobilization readiness ≠ stage readiness
+
+- File 04 = "if we got approval today, could the crew start?"
+- `@Kushi fde-report <project> stage-readiness` = "should we advance to the next FDE stage?"
+
+Run both for a stage gate.
+
+## Re-runs
+
+Re-running on the same day overwrites files 00–06 in place and **merges file 07** (preserves user-set `resolved` / `not-applicable` statuses).
+
+## See also
+
+- `skills/fde-triage/SKILL.md` for the full step-by-step.
+- `@Kushi fde-intake <project>` — first-artifact intake document (precursor to the bundle).
+- `@Kushi fde-report <project> [shape]` — individual report shapes (weekly / short / long / fitness / stage-readiness).

package/plugin/prompts/refresh.prompt.md

@@ -0,0 +1,17 @@
+---
+name: refresh
+description: Incremental refresh — pull only what changed since the last run, then rebuild State.
+---
+
+# /refresh
+
+Route to `@Kushi refresh <project>`.
+
+Window resolution (per `refresh-project` skill):
+1. If user gave explicit window (`last 7 days`, `since <date>`, `<from>..<to>`) → use it.
+2. Else read `Evidence/run-log.yml` → use `min(sources.*.watermark)` as `<from>`.
+3. Else fallback to last 7 days.
+
+Output is always weekly-bucketed regardless of input window.
+
+Delegates to `refresh-project` skill. Skips `pull-<source>` for sources marked disabled in `integrations.yml` or whose config is missing.

package/plugin/prompts/state.prompt.md

@@ -0,0 +1,17 @@
+---
+name: state
+description: Re-render State/ from existing Evidence — does NO source pulls.
+---
+
+# /state
+
+Route to `@Kushi state <project>`.
+
+Pure renderer. Reads:
+- `<engagement-root>/<project>/Evidence/<alias>/**` (all snapshot + stream files for the user)
+- `<engagement-root>/<project>/Evidence/_Consolidated/**` if present
+
+Writes:
+- `<engagement-root>/<project>/State/{00_overview, 01_decisions, 02_stakeholders, 03_architecture-and-solution, 04_workshops-and-key-meetings, 05_action-items, 06_risks-and-issues, 07_timeline-and-milestones, 08_artifacts-and-deliverables, 09_open-questions}.md`
+
+Delegates to `build-state` skill. Use this when you've manually edited Evidence files and want State refreshed without re-pulling.

package/plugin/prompts/status.prompt.md

@@ -0,0 +1,17 @@
+---
+name: status
+description: Show run-log — what's been pulled, when, watermarks, recent errors.
+---
+
+# /status
+
+Route to `@Kushi status <project>`.
+
+Reads `<engagement-root>/<project>/Evidence/run-log.yml` and renders:
+- Per-source: last_pulled, watermark, item_count
+- Recent weekly_files index
+- Any pull errors logged
+
+Read-only. No pulls, no writes.
+
+Delegates to `project-status` skill.

package/plugin/reference-packs/README.md

@@ -0,0 +1,74 @@
+# Kushi reference packs
+
+This folder is the **canonical source for static reference content** Kushi skills consult at runtime — domain operating models, taxonomies, stage gates, criteria, status-meaning tables, and similar mostly-stable doctrine.
+
+Each subdirectory under `reference-packs/` is one **pack**. Packs are bundled with the Kushi npm package, versioned in git, and installed onto consumer machines via the `full` install profile (or any profile that opts in to the pack).
+
+## How packs are loaded
+
+When a skill (e.g. `fde-report`, `ask-project`) needs a pack's content, it loads files **in this priority order**, highest wins:
+
+| Priority | Location | Purpose | Lifecycle |
+|---:|---|---|---|
+| 1 (highest) | `<engagement-root>/<project>/.kushi-reference/<pack>/` | Per-engagement override | Versioned with the engagement; for project-specific tuning |
+| 2 | `~/.copilot/kushi-reference/<pack>/` | Per-user override | Per machine / per user; never overwritten by installer |
+| 3 (canonical) | `<install-dest>/reference-packs/<pack>/` (i.e. this folder, installed) | Packaged default | Updated via `npm publish` + version bump |
+
+Skills must always:
+
+1. Read the pack's `README.md` first (it documents read-order, file ownership, and any conflict-resolution rules specific to that pack).
+2. Enumerate **all human-readable `.md` files in the pack** (default canonical first, then merge in user override, then project override).
+3. Treat the union as one reference set — but when files conflict, prefer the higher-priority layer and call out the override in the skill's run-log / Open Questions instead of silently averaging.
+
+## What goes in a pack
+
+- Operating-model definitions (what good looks like; anti-patterns)
+- Fitness criteria, gates, exit criteria, checklists
+- Status-meaning tables (e.g. CRM lifecycle codes)
+- Risk taxonomies
+- Reporting guardrails specific to a domain
+- "Spotlight" reference cases / canonical examples
+- Customer-readiness criteria
+
+What does **not** go in a pack:
+
+- Per-engagement data (lives under `<project>/Evidence/`)
+- Per-user identity / configs (lives in `~/.copilot/project-evidence.yml`)
+- Skill behavior (lives in `plugin/skills/<skill>/SKILL.md`)
+- Cross-cutting agent doctrine (lives in `plugin/instructions/*.instructions.md`)
+
+## Pack structure (required)
+
+```
+reference-packs/<pack-name>/
+├── README.md                      ← required; documents read behavior + file ownership
+├── <topic-1>.md                   primary reference file(s)
+├── <topic-2>.md
+└── ...
+```
+
+`README.md` must explain:
+
+- What the pack covers
+- Which skills consume it
+- Read order (which file is the canonical source for which concept)
+- How conflicts between files should be resolved
+
+## Current packs
+
+| Pack | Profile | Consumed by | Status |
+|---|---|---|---|
+| `fde/` | `full` | `fde-report`, `ask-project` (for FDE-shaped questions) | shipped |
+
+## Adding a new pack
+
+1. Create `plugin/reference-packs/<pack-name>/` with at minimum a `README.md` describing read behavior.
+2. Add the pack name to one or more profiles in `plugin/plugin.json` under `reference_packs: ["<pack-name>", ...]`.
+3. Reference the pack from the consuming skill's `SKILL.md`, using the [3-layer load order](#how-packs-are-loaded).
+4. Run `pwsh plugin/skills/self-check/run.ps1 -Deep` — check **C10** validates that every pack referenced by a profile exists and has a `README.md`.
+
+## See also
+
+- [Concepts → Reference packs](https://gim-home.github.io/kushi/concepts/reference-packs/) — the override model in plain English
+- [Reference → Reference packs](https://gim-home.github.io/kushi/reference/reference-packs/) — schema, load order, override semantics
+- [Concepts → Profiles and report packs](https://gim-home.github.io/kushi/concepts/profiles-and-packs/) — which profile ships which packs

package/plugin/reference-packs/fde/README.md

@@ -0,0 +1,62 @@
+# FDE Reference Pack
+
+This pack is the canonical home for **Forward Deployed Engineering (FDE)** reporting guidance used by stage-readiness, FDE-fit, status, and risk outputs produced by Kushi.
+
+Use this pack when a report needs:
+
+- FDE definitions and anti-patterns
+- FDE fitness criteria
+- CRM status → FDE-stage interpretation
+- Stage definitions, stage gates, checklists, exit criteria, and deliverables
+- FDE-specific risk categories and readiness framing
+
+## Consumed by
+
+- **`fde-intake`** — generates the engagement's first artifact (`00-FDE-Intake-<project>.md`). Reads `intake-questions.md` + `core-fde-reference.md`.
+- **`fde-report`** — single-shape report generator (`weekly` / `short` / `long` / `fitness` / `stage-readiness`). Reads all three pack files.
+- **`fde-triage`** — 7-file triage bundle generator. Reads all three pack files; relies heavily on `report-doctrine.md` for the validation warnings protocol.
+- **`ask-project`** — when a question maps to FDE concepts (e.g. "what stage is project X in?", "what's still pending to advance to Assigned?", "is this an FDE-fit engagement?"), pulls relevant files from this pack as additional grounding alongside the project's own Evidence/.
+
+## Read behavior
+
+1. **Read `README.md` first** (this file).
+2. **Enumerate and read all human-readable `.md` files** in the pack.
+3. Treat the pack as a **single reference set** — any file here may contribute to the FDE operating model used in reports.
+4. **Apply the 3-layer override order** (highest wins):
+   - `<engagement-root>/<project>/.kushi-reference/fde/`  ← per-engagement override
+   - `~/.copilot/kushi-reference/fde/`                   ← per-user override
+   - `<install-dest>/reference-packs/fde/`               ← packaged default (this folder, installed)
+5. If two files at the same layer conflict, prefer the more **specific** file for that concept and raise the conflict explicitly in the skill's run-log / Open Questions — never silently average.
+
+## File ownership
+
+- **`core-fde-reference.md`** — canonical source for the **FDE operating model** (definitions, anti-patterns, fitness criteria, delivery model, CRM status meanings, value proposition, risk categories, customer readiness criteria, intake stages and gates).
+- **`report-doctrine.md`** — canonical source for FDE **report rules** (source resolution gate, recency precedence, CRM-field-vs-confirmed-fact, validation warnings protocol with Rule X.Y numbering, citation density, privacy posture for external-facing reports).
+- **`intake-questions.md`** — canonical source for the **6 FDE Intake questions + required/optional role rules** (ATU + Industry Team required; ISE / ISD / Exec Sponsor / PG optional with explicit "absent" statements).
+
+If this pack grows, each concept should still have a single owning file:
+
+- Do not split the same stage gate, readiness rule, or risk taxonomy across multiple files unless one file is explicitly the *summary* and another is explicitly the *detail*.
+- Skills should consult every file but route citations to the owning file.
+
+## Canonical location
+
+`plugin/reference-packs/fde/` in this repository.
+
+After install: `<install-dest>/reference-packs/fde/` (i.e. `.kushi/reference-packs/fde/` for VS Code, `~/.copilot/m-skills/kushi/reference-packs/fde/` for Clawpilot).
+
+## Citations from this pack
+
+When `fde-report` or `ask-project` quotes from this pack in an output, the citation form is:
+
+```text
+[source: reference-packs/fde/<file>.md · packaged]
+```
+
+`packaged` denotes the canonical layer (i.e. this folder shipped in the npm package). Overrides cite as `user-override` or `project-override` respectively, so the reader can tell what layer the assertion came from.
+
+## Editing
+
+Edits to the canonical files in this folder ship as a **new Kushi minor release** (`v3.x.0` → `v3.(x+1).0`). For one-off tweaks specific to your machine or one engagement, prefer the **user override** or **project override** layers above — those never get overwritten by re-installs.
+
+See [`plugin/reference-packs/README.md`](../README.md) for the full pack-system doctrine.