@gobing-ai/spur 0.3.92 → 0.3.94

package/plugins/sp/commands/dev-review.md

@@ -1,7 +1,7 @@
 ---
-description: "Review code for a task or path — multi-dimensional review across functional traceability, SECUA quality, and architectural depth. Triggers: \"review this\", \"check the code\", \"SECUA review\", \"dev review\", \"audit this\"."
+description: "Review code for a task or path — multi-dimensional review across functional traceability, SECUA quality, and architectural depth. With --triage: fix small findings directly, then file the rest as implement-ready tasks. Triggers: \"review this\", \"check the code\", \"SECUA review\", \"dev review\", \"audit this\", \"triage review findings\"."
 role: reviewer
-argument-hint: "[<wbs|path>] [--agent <inline|auto|name>] [--focus <dims>] [--fix (deprecated)]"
+argument-hint: "[<wbs|path>] [--agent <inline|auto|name>] [--focus <dims>] [--triage] [--worktree [<name>]]"
 allowed-tools: ["Bash", "Read", "Skill"]
 ---
 
@@ -16,17 +16,28 @@ Wraps the **sp:functional-review**, **sp:code-verification**, and **sp:code-impr
 | `[<wbs\|path>]` | Task WBS or source path to review. | cwd |
 | `--agent` `<inline\|auto\|name>` | Who runs the model-bearing review. | omit |
 | `--focus` `<dims>` | Review dimensions (functional / SECUA / architecture). | all |
-| `--fix` | Deprecated no-op + warning; route remediation to /sp:dev-verify --fix. | off |
+| `--triage` | After the review: fix small findings directly, file the rest as one or more implement-ready tasks. | off (report-only) |
+| `--worktree` `[<name>]` | Run the triage in an isolated git worktree; FF-merge on success, retain on failure. Bare `--worktree` creates a fresh tree; `--worktree <name>` adopts an existing worktree by name/path/branch. Requires `--triage`. | off |
 
 For shared semantics, see the [flag glossary](../skills/spur-dev/references/flag-glossary.md).
 
 ## Usage
 
-/sp:dev-review [<wbs|path>] [--agent <inline|auto|name>] [--focus <dims>] [--fix (deprecated)]
+```
+/sp:dev-review [<wbs|path>] [--agent <inline|auto|name>] [--focus <dims>] [--triage] [--worktree [<name>]]
+/sp:dev-review scripts --triage
+/sp:dev-review 0962 --triage --worktree
+```
+
+Put the target first: a bare token after `--worktree` is read as the worktree name.
 
 ## Implementation
 
 - Apply the [inline-default execution-surface contract](../skills/spur-dev/references/cross-cutting.md#inline-default-execution-surface).
 - WBS mode (`<wbs>`): `Skill(skill="sp:functional-review", args="<wbs> $ARGUMENTS")` + `Skill(skill="sp:code-verification", args="review $ARGUMENTS")` + `Skill(skill="sp:code-improvement", args="<wbs> $ARGUMENTS")` (functional traceability + SECUA framework + architectural depth; the three skills return review fragments and the coordinator (`sp:super-reviewer`) writes the combined `## Review` — F92 0593 R1)
 - Path mode (`<path>`): `Skill(skill="sp:code-verification", args="review $ARGUMENTS")` + `Skill(skill="sp:code-improvement", args="<path> $ARGUMENTS")` (advisory SECUA quality + architectural depth; performs no task mutation)
-- `--fix`: Deprecated (no-op + warning message; route remediation to `/sp:dev-verify --fix`). **`--next` removed** (feature H8, 2026-07-31) — it was a deprecated no-op; route progression to `/sp:dev-next`. **was: `--next` deprecated no-op.**
+- `--triage`: after the review returns, run the triage protocol in [dev-operations.md § 2. review](../skills/spur-dev/references/dev-operations.md#2-review) **inline in this session** — the review skills and `sp:super-reviewer` stay report-only.
+- Without `--worktree`, every write lands in the current working tree on the current branch: no branch creation, no checkout, no commit.
+- `--worktree [<name>]`: the [execution-batch.md § Worktree isolation](../skills/spur-dev/references/execution-batch.md#worktree-isolation---worktree-name) lifecycle as a run of one (marker `command` = `dev-review`, `selector` = the target). Rejected without `--triage`.
+  While it runs, fixes and filed tasks land in the worktree copy; the main tree shows them only after the FF-merge.
+- `--fix`: Deprecated (still accepted: no-op + warning; route remediation to `--triage` or `/sp:dev-verify --fix`). **`--next` removed** (feature H8, 2026-07-31) — it was a deprecated no-op; route progression to `/sp:dev-next`. **was: `--next` deprecated no-op.**

package/plugins/sp/commands/dev-run.md

@@ -15,7 +15,7 @@ Wraps the **sp:spur-dev** and **sp:code-implementation** skills.
 | --- | --- | --- |
 | `<wbs>` | Task WBS to run. | required |
 | `--mode` `<full\|implement>` | Full pipeline or single implement step. | full |
-| `--agent` `<inline\|auto\|name>` | Who runs the model-bearing stages. omit and explicit `--agent inline` resolve identically (task 0687): eligible stages dispatch once to a native subagent with host-session fallback (0508 eligibility); otherwise every stage executes in the invoking session. `auto` or a name keeps subprocess dispatch.. | omit |
+| `--agent` `<inline\|auto\|name>` | Who runs the model-bearing stages. omit and explicit `--agent inline` resolve identically (task 0687): eligible stages dispatch once to a native subagent with host-session fallback (0508 eligibility); otherwise every stage executes in the invoking session. `auto` or a name keeps subprocess dispatch.. The opt-in `fleet` selector (0942) maps the run to `executor: 'fleet'` so `agent.run` stages dispatch through the fleet control plane (ADR-126) instead of a subprocess. | omit |
 | `--auto` | Skip objective HITL confirmations. | off |
 | `--next` | Chain-to-completion via the next-router. | off |
 | `--wrap` | Run the wrap hop after the main step. The `--agent` selector is preserved into the `/sp:dev-wrap <wbs>` handoff when supplied; omission remains omission. The wrap hop is workflow-backed and reports its trigger-3 subprocess override. | off |
@@ -40,7 +40,7 @@ For shared semantics, see the [flag glossary](../skills/spur-dev/references/flag
 
 **Flags:**
 
-- `--auto` | `--agent <inline|auto|name>` — Skip objective HITL confirmations (taste/irreversible gates still pause). `--agent` names who does the model-bearing work. Interactive omit/`inline` keeps the controller and implement-only stages in this session; full mode reads `task-pipeline.yaml` as the SSOT and interprets its actions/guards through the inline driver, whose eligible `agent.run` stages dispatch once to a native subagent and otherwise run in the host (0508 eligibility — task 0687 resolved-inline) It records `stage <id> executed inline in session <session-id>` or `stage <id> executed via subagent <agent-id> (host session <session-id>)` in the run log. `auto` or a name is merged into `vars.agent` and `vars.implementAgent` and keeps the existing subprocess workflow. Headless `spur workflow run` / `spur agent run` is unchanged. See the [execution-surface contract](../skills/spur-dev/references/cross-cutting.md#inline-default-execution-surface).
+- `--auto` | `--agent <inline|auto|name>` — Skip objective HITL confirmations (taste/irreversible gates still pause). `--agent` names who does the model-bearing work. Interactive omit/`inline` keeps the controller and implement-only stages in this session; full mode reads `task-pipeline.yaml` as the SSOT and interprets its actions/guards through the inline driver, whose eligible `agent.run` stages dispatch once to a native subagent and otherwise run in the host (0508 eligibility — task 0687 resolved-inline) It records `stage <id> executed inline in session <session-id>` or `stage <id> executed via subagent <agent-id> (host session <session-id>)` in the run log. `auto` or a name is merged into `vars.agent` and `vars.implementAgent` and keeps the existing subprocess workflow; `--agent fleet` maps the run to `vars.executor: 'fleet'` instead (fleet dispatch, 0942). Headless `spur workflow run` / `spur agent run` is unchanged. See the [execution-surface contract](../skills/spur-dev/references/cross-cutting.md#inline-default-execution-surface).
 - `--escalation-file <path>` (implement step only, 0933) — names the Q/A escalation transcript
   (default `.spur/run/<wbs>-escalation.md`). Absent on a first attempt means "no prior
   escalations". When the implement agent needs an operator decision it appends its question to

package/plugins/sp/commands/dev-runall.md

@@ -18,7 +18,7 @@ Wraps the **sp:spur-dev** skill.
 | `--mode` `<sequential\|parallel>` | Batch execution order. | sequential |
 | `--keep-going` | Continue past per-task failures. | off |
 | `--auto` | Skip objective HITL gates. | off |
-| `--agent` `<inline\|auto\|name>` | Who runs each task's pipeline stages. Interactive sequential omit/`inline` uses the host-session driver with unified inline semantics (task 0687 — see below)). `auto`, a name, parallel mode, and headless invocation use subprocesses. | omit |
+| `--agent` `<inline\|auto\|name>` | Who runs each task's pipeline stages. Interactive sequential omit/`inline` uses the host-session driver with unified inline semantics (task 0687 — see below)). `auto`, a name, parallel mode, and headless invocation use subprocesses. The opt-in `fleet` selector (0942) maps each task run to `executor: 'fleet'` so `agent.run` stages dispatch through the fleet control plane (ADR-126) instead of a subprocess. | omit |
 | `--json` | Emit structured JSON. | off |
 | `--wrap` | Run the wrap hop **once for the batch** over the `done` subset only (Step 6 of execution-batch.md): `vars.feature` is passed only when every frozen task is `done`/`cancelled`; an empty done subset skips the wrap with a reason. The `--agent` selector is preserved into the `/sp:dev-wrap` handoff when supplied; omission remains omission. | off |
 | `--next` | Chain-to-completion via the next-router. | off |
@@ -49,7 +49,7 @@ only after its in-set dependencies are **integrated** onto the base ref), `--kee
 halts on first failure), `--auto`
 (sets `profile=auto` on each per-task run, skipping the HITL approve gate), `--agent <inline|auto|name>`
 (names who runs the pipeline's stages — interactive sequential omit/`inline` uses the host driver;
-`auto` or a name is merged into each per-task `vars.agent` and `vars.implementAgent`; parallel mode
+`auto` or a name is merged into each per-task `vars.agent` and `vars.implementAgent`; `fleet` maps each task run to `executor: 'fleet'` instead (fleet dispatch, 0942); parallel mode
 retains isolated subprocesses, ADR-047), `--json` (emit the
 report as JSON), `--wrap` (trigger
 `wrapup-pipeline.yaml` after the batch completes), `--next`

package/plugins/sp/hooks/pi/guard-extension.ts

@@ -25,6 +25,7 @@ import { isAbsolute, join, resolve } from 'node:path';
 import type { ExtensionAPI } from '@earendil-works/pi-coding-agent';
 import { getEnvVar, getEnvVars } from '../../lib/env';
 import { resolveAgentHint as resolveAgentHintShared, resolveModelHint as resolveModelHintShared } from '../agent-hint';
+import { cappedByteLength, truncateSummary } from '../context-post-tool';
 import { classifyCommand } from '../destructive-policy';
 import { couldBeTaskFile } from '../task-file-policy';
 
@@ -36,8 +37,6 @@ import { couldBeTaskFile } from '../task-file-policy';
 const spurContextDir = (): string => join(process.cwd(), '.spur', 'context');
 const sessionFilePath = (): string => join(spurContextDir(), '.session.json');
 const ledgerFilePath = (): string => join(spurContextDir(), 'token-ledger.jsonl');
-const REDACTION_CAP = 4096;
-const SUMMARY_MAX_CHARS = 200;
 
 // ─── Helpers ─────────────────────────────────────────────────────────────
 
@@ -122,39 +121,17 @@ interface ToolEvent {
     tool_input?: Record<string, unknown>;
 }
 
-function estimateTokenCount(text: string): number {
-    if (!text) return 0;
-    // Rough estimate: ~4 chars per token for English text
-    return Math.ceil(text.length / 4);
-}
-
+// Summaries persist to disk, so they go through the same scrubber as the Claude hook —
+// the local copy this replaced stored the raw command (secrets included) as `summary`.
 function summarizeToolEvent(event: ToolEvent): string {
     const input = event.tool_input ?? {};
     const candidates = [input.file_path, input.command, input.pattern, input.glob_pattern, input.glob, input.path];
     for (const c of candidates) {
-        if (typeof c === 'string' && c.trim()) {
-            const s = c.trim();
-            return s.length > SUMMARY_MAX_CHARS ? `${s.slice(0, SUMMARY_MAX_CHARS - 3)}...` : s;
-        }
+        if (typeof c === 'string' && c.trim()) return truncateSummary(c);
     }
     return `(${event.tool_name ?? 'unknown'})`;
 }
 
-function redactText(text: string): string {
-    // Strip obvious secret patterns
-    let result = text.slice(0, REDACTION_CAP);
-    result = result.replace(
-        /-----BEGIN\s+(?:RSA\s+)?PRIVATE\s+KEY-----[\s\S]*?-----END\s+(?:RSA\s+)?PRIVATE\s+KEY-----/g,
-        '[REDACTED: PRIVATE KEY]',
-    );
-    result = result.replace(/ghp_[A-Za-z0-9]{36}/g, '[REDACTED: GITHUB_TOKEN]');
-    result = result.replace(/gho_[A-Za-z0-9]{36}/g, '[REDACTED: GITHUB_TOKEN]');
-    result = result.replace(/sk-[A-Za-z0-9]{20,}/g, '[REDACTED: API_KEY]');
-    result = result.replace(/api[_-]key['"]?\s*[:=]\s*['"][A-Za-z0-9_-]{16,}['"]/gi, '[REDACTED: API_KEY]');
-    result = result.replace(/AKIA[0-9A-Z]{16}/g, '[REDACTED: AWS_KEY]');
-    return result;
-}
-
 function appendToLedger(event: ToolEvent, command: string | undefined): void {
     try {
         if (!existsSync(spurContextDir())) return;
@@ -162,16 +139,19 @@ function appendToLedger(event: ToolEvent, command: string | undefined): void {
         if (!sessionId) return;
 
         const summary = summarizeToolEvent(event);
-        const tokens = command ? estimateTokenCount(redactText(command)) : 0;
+        const tokens = command ? Math.ceil(cappedByteLength(command) / 4) : 0;
+        const input = event.tool_input ?? {};
+        // Pi tool names are lowercase (`read`) and its write/edit tools take `path`, not `file_path`.
+        const path = typeof input.file_path === 'string' ? input.file_path : input.path;
 
         const ledgerEntry = JSON.stringify({
             session: sessionId,
-            type: event.tool_name === 'Read' ? 'read' : 'write',
+            type: event.tool_name?.toLowerCase() === 'read' ? 'read' : 'write',
             tool: event.tool_name,
-            path: event.tool_input?.file_path ?? null,
+            file: typeof path === 'string' ? path : undefined,
             summary,
             tokens,
-            timestamp: new Date().toISOString(),
+            ts: new Date().toISOString(),
         });
 
         appendFileSync(ledgerFilePath(), `${ledgerEntry}\n`);
@@ -216,7 +196,7 @@ function initSession(): void {
             type: 'session_start',
             agent: session.agent,
             model: session.model,
-            timestamp: session.started_at,
+            ts: session.started_at,
         });
         appendFileSync(ledgerFilePath(), `${startEntry}\n`);
     } catch {
@@ -256,7 +236,7 @@ function cleanupSession(): void {
             reads,
             writes,
             tokens,
-            timestamp: new Date().toISOString(),
+            ts: new Date().toISOString(),
         });
         appendFileSync(ledgerFilePath(), `${endEntry}\n`);
 
@@ -272,8 +252,9 @@ function cleanupSession(): void {
 export default function (pi: ExtensionAPI): void {
     // ── task-write-guard + careful-guard ──────────────────────────────
     pi.on('tool_call', async (event, ctx) => {
-        // task-write-guard: block Write/Edit to Spur task files
-        if (event.toolName === 'write' || event.toolName === 'edit') {
+        // task-write-guard: block Write/Edit to Spur task files.
+        // Both guards honor the same `off` escape hatches as the Claude hooks.
+        if ((event.toolName === 'write' || event.toolName === 'edit') && getEnvVar('SPUR_WRITE_GUARD') !== 'off') {
             const input = event.input as Record<string, unknown> | undefined;
             // Pi's write/edit tools use `path` (not Claude Code's `file_path`)
             const filePath = resolveInputPath(input);
@@ -287,7 +268,7 @@ export default function (pi: ExtensionAPI): void {
         }
 
         // careful-guard: warn on destructive Bash commands
-        if (event.toolName === 'bash') {
+        if (event.toolName === 'bash' && getEnvVar('SPUR_CAREFUL') !== 'off') {
             const input = event.input as Record<string, unknown> | undefined;
             const command = typeof input?.command === 'string' ? input.command : '';
             const hit = command ? classifyCommand(command) : null;