@tacuchi/agent-workflow-cli 20.12.0 → 20.13.0

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@tacuchi/agent-workflow-cli",
-  "version": "20.12.0",
+  "version": "20.13.0",
   "description": "Runtime CLI for Workline — the stages + loops + artifacts system for agent work. Bundles the universal `w` skill set under `skills/w/` (slash commands `/w:*`: spec-new/spec-refine, plan-new/plan-exec, quick, persist, workspace-init, export-*); `self install --target <host>` copies SKILL + commands + hooks into the host. Pluggable capability skills via `.workflow/skills.toml`. Multi-empresa parametrization via `profile.json` cascade. Namespace auto-detected from any `.<ns>/sessions/` dir in CWD; default `workflow`.",
   "type": "module",
   "bin": {

package/skills/w/harness/HARNESS.md

@@ -39,6 +39,7 @@ The capabilities the harness layer depends on, with their universal fallback (wh
 | **subagent-dispatch** | *(optional)* parallelize research breadth | **inline sequential** research in the same session (the default anyway) |
 | **persistent-context** | the `WORKSPACE` block + conventions always present | the repo's context file (standard **`AGENTS.md`**; `CLAUDE.md` on Claude Code) |
 | **host-memory** | *(optional)* recover state/pending work from the host's accessible history — a **second source** after the workline signals | recent **git** / **`docs/`** signals + (in `/resume`) **ask the user**; plus Workline's own `.workflow/CHECKPOINT` via `aw resume-summary` |
+| **web-research** | *(optional)* search/fetch external online evidence inside a consented ideation round (spec-refine § *Ideation gate*) | **offline ideation** (own knowledge + workspace + repos) — the loop **declares** the web was unavailable |
 | **external-data** | read-only DB reads or other sources for research/validation | **MCP** (widely supported); without it, the gap degrades to a human question |
 | **dry-run / preview** | preview what a command would do without writing | the command **describes** the change instead of applying it (e.g. `spec-new` lists the draft without creating the file) |
 
@@ -57,6 +58,7 @@ Concrete mechanism per harness (**Jul-2026**, verified against official docs; `~
 | subagent-dispatch | `Task` (parallel) | `SubagentStart` / agents | agents (`.gemini/agents`) | `.opencode/agent/*.md` | ~ | ~ (cloud agents) | inline |
 | persistent-context | `CLAUDE.md` (does **not** read AGENTS.md → symlink) | `AGENTS.md` | `GEMINI.md` + `AGENTS.md` | `AGENTS.md` | `CRUSH.md` + `AGENTS.md` | `AGENTS.md` (auto) | `AGENTS.md` |
 | **host-memory** | `MEMORY.md` (cheap) + transcripts/`--resume` (deep) | `AGENTS.md` (static → fallback) | `GEMINI.md`+`AGENTS.md` (static → fallback) | `AGENTS.md` (static → fallback) | `CRUSH.md`+`AGENTS.md` (static → fallback) | rules / history (~) | git/`docs/` + ask |
+| **web-research** | `WebSearch` / `WebFetch` | `web_search` (opt-in config) | `google_web_search` + `web_fetch` | `webfetch` (~) | ~ | ~ (agent web access) | — (offline + declare) |
 | external-data (MCP) | `.mcp.json` | `.codex/config.toml` `[mcp_servers]` | `settings.json` `mcpServers` | `opencode.json` `mcp` | `crush.json` `mcp` | `.warp/.mcp.json` (+auto-discovers `.mcp.json`) · Oz: `--mcp` flag | — |
 | **enforcement (deny tool)** | `PreToolUse` → `permissionDecision:deny` / exit 2 | `PreToolUse` (**≈same protocol**) | `BeforeTool` → `decision:deny` / exit 2 | plugin `tool.execute.before` (`throw`) | `allowed_tools` (+ preliminary hooks) | allow/deny lists (**coarse**) | doctrine (git-safe #5) |
 | plugin / dist | `.claude-plugin` + marketplace | `.codex-plugin` + `/plugins` marketplace | Extension `gemini-extension.json` | JS/TS plugin (npm) | MCP + skills + config | Warp Drive | — |
@@ -69,6 +71,8 @@ Concrete mechanism per harness (**Jul-2026**, verified against official docs; `~
 
 > **host-memory (tiers & consumers).** Two tiers: *cheap* (structured, bounded — on Claude Code the auto-memory `MEMORY.md` + `CLAUDE.md`) and *deep* (transcript / `--resume` search, expensive). Consumers: **`/status`** reads only the *cheap* tier, **opportunistically and additively** (a `CONTEXTO DEL HOST` section when available; it **never asks** — a read-only dashboard — and silently omits the section on degrade); **`/resume`** **composes `/status`** and escalates a host-only finding **to a proposal only when the workline level does not explain the pending work** (the spec's fixed order governs the proposals, not the summary), optionally using the *deep* tier or asking as fallback. It is *enhancement*, never a `must`.
 
+> **web-research (consumer & consent).** Single consumer today: `spec-refine-loop` § *Ideation gate* (the SPEC flow's divergent gate). Consent is **bundled into accepting that gate's offer** — one accepted offer authorizes that round's searches; never free browsing outside a consented round. It is *enhancement*, never a `must`: on degrade the loop ideates **offline and declares it** (unlike host-memory's silent omit — the user consented expecting the web, so silence would mislead).
+
 ## Leverage installed skills
 
 "Leverage whatever skills the harness has installed" resolves through the **same** `.workflow/skills.toml` binding: a role can point at a skill **installed on the host** (third-party, via skills.sh) instead of the built-in. Rule:

package/skills/w/loops/spec-refine-loop/LOOP.md

@@ -3,7 +3,8 @@ name: spec-refine-loop
 description: >-
   Refines a draft spec (docs/specs/NNN-spec-<slug>.md) by editing it IN PLACE
   until it is unambiguous. Heir of the chassis (loops/CHASSIS.md). Deltas:
-  spec gap taxonomy, analyze gate, ## UI spec section via the ui-design
+  spec gap taxonomy, analyze gate, ideation gate (consented brainstorming,
+  optional web-research), ## UI spec section via the ui-design
   capability, and adds Refinement decisions + Q&A traceability — the refined
   mark plan-new detects. Started by /w:spec-refine (or the live escalation
   from quick-loop); resumable via CHECKPOINT and re-runnable on demand.
@@ -81,7 +82,7 @@ Structured Markdown description (screens → regions/components). See [`ui-spec`
 
 ## Refinement decisions   ← NEW (ADDED)
 What was defined while refining and why. Includes what inline research
-resolved (with a reference to the session's CONCLUSIONS).
+resolved (ref. the session's CONCLUSIONS) and the ideation verdicts (§ Ideation gate).
 
 ## Q&A traceability       ← NEW (ADDED)
 Every doubt asked to the human + the chosen answer.
@@ -108,8 +109,18 @@ Every doubt asked to the human + the chosen answer.
 | Hidden assumptions | the spec assumes unstated things | **research** validates / **human** confirms |
 | Internal contradiction | sections contradict each other | **human** |
 | Over-specified requirement | scope/criteria gold-plated — beyond the actual need (chassis § *Minimality*) | **human** (AI proposes the cut, human ratifies) |
+| Unexplored solution space | the spec settles on the first conceivable approach — no alternatives weighed nor adopted from the conversation | **human consents** → **ideation** (see *Ideation gate*) |
 | UI unspecified *(if it applies)* | the requirement involves UI but `## UI spec` is missing | **`ui-design` capability** |
 
+## Ideation gate (creativity)
+
+The loop's one **divergent** gate: every other resolver closes a gap; this one widens the option space before the spec hardens around its first idea. It resolves **Unexplored solution space** in two consented steps:
+
+1. **Offer & consent.** The gap enters the batch as a content question — `Explorar ideas` vs `Seguir sin ideación` — carrying the AI's recommendation like any other. Declining marks the gap **exhausted** (never re-offered this run); an explicit user request for ideas at any point counts as an accepted offer (on-demand entry). Alternatives already weighed in the conversation are *adopted context* — the gap does not fire.
+2. **Ideation round** (one per consent). Propose fresh ideas and **combinations** (the user's + found ones). If the host exposes **web-research** ([`../../harness/HARNESS.md`](../../harness/HARNESS.md)), the accepted offer also authorizes that round's web searches — no per-search consent; findings + sources land in the session's `CONCLUSIONS`, like inline research. Without the capability, ideate offline (own knowledge + workspace + repos) and **declare it** — never silently.
+
+**Verdicts (back to convergence).** Present the top ≤3 ideas via the same structured-choice, each with a recommended verdict: `Adoptar` → integrate into `Requirement`/`Scope`/criteria + trace in `## Refinement decisions` (idea, rationale, source/URL when web-found) · `Descartar` → one line + reason there · `Aparcar` → `## Open questions`. Ideas beyond the top 3 stay summarized in `CONCLUSIONS`. Divergence is bounded by *Minimality* (chassis): nothing enters the spec without an explicit `Adoptar`; the analyze gate keeps pruning gold-plating. This gate exists **only** in this loop — `spec-new` stays single-pass (no research, no web) and the plan/quick loops inherit none of it.
+
 ## Sequence
 
 ```
@@ -128,6 +139,8 @@ spec-refine-loop(spec):
       if gap = UI (requirement involves UI, ## UI spec missing):
         compose ui-design → author ## UI spec    # design-system/theme via structured-choice (counts in the batch)
         work = integrate(work, ui)               # → ## UI spec
+      else if gap = Unexplored solution space (creativity):
+        pending_human.push("ideation offer")     # offer with AI recommendation: `Explorar ideas` | `Seguir sin ideación`
       else if factual(gap) and attempts[gap] < MAX:
         if it needs DB and >1 MCP without default → queue "MCP choice" in pending_human
         res = research_inline(gap)           # in the current session: ANALYSIS-FILE → CONCLUSIONS (+read-only SCRIPTS.sql)
@@ -142,6 +155,8 @@ spec-refine-loop(spec):
         Compactar → write CHECKPOINT (refine_session) ; compact(harness) ; continue
         Cerrar    → goto finalize
       work = integrate(work, ans)            # → Q&A traceability / Open questions
+      ideation offer accepted → run the round NOW, then its verdicts as a NEW ≤3+flow batch (§ Ideation gate) → integrate
+      ideation offer declined → mark that gap exhausted    # anti re-fire; on-demand entry stays open
   # no material gaps → analyze gate = Success criteria green (read-only) before offering Guardar:
   issues = analyze(work)   # criteria trace to the Requirement · no contradictions · coherent Scope · Open questions closed/deferred · scenarios↔criteria · no gold-plating (minimality)
   if issues: gaps += issues ; continue            # findings come back into the loop as gaps
@@ -172,6 +187,7 @@ Full mechanism (3 cases, `Compactar`, re-run on demand with `--reopen`) in the c
 ## Integration (where each resolution lands)
 
 - Resolved via **inline research** → the spec's `## Refinement decisions` (+ ref to the session's `CONCLUSIONS`).
+- Resolved via **ideation** → per verdict (§ *Ideation gate*): `Adoptar` → the spec's sections + `## Refinement decisions` · `Descartar` → there · `Aparcar` → `## Open questions`.
 - Resolved via **human** → the spec's `## Q&A traceability`.
 - Resolved via the **`ui-design` capability** (UI gap) → the spec's `## UI spec` section.
 - **Inconclusive or unresolved research** → the spec's `## Open questions` (deferred) + the refine session's `BACKLOG.md` (only if something is deferred).