svamp-cli 0.2.346 → 0.2.356

package/bin/skills/artifact/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: artifact
-version: 1.3.1
+version: 1.3.2
 description: Embed a rendered HTML file or live URL inline in chat via a single self-closing `<artifact src="..." />` tag — the Svamp client renders it as a sandboxed iframe with modes for inline, card, bare, and immersive.
 tags:
   - svamp
@@ -134,16 +134,30 @@ This means you can iterate freely without polluting the chat with stale renders.
 To share with a colleague, run `svamp serve`:
 
 ```bash
-# Owner-only (default — they need to log in via Hypha)
+# DEFAULT is the `link` tier — a capability URL. NO login required:
+# anyone who has the URL can read it. The hostname IS the secret.
 svamp serve dashboard ./outputs
 
-# Public link (anyone with the URL)
-svamp serve dashboard ./outputs --public
+# Login-gated (only you, via Hypha) — opt IN to this
+svamp serve dashboard ./outputs --access owner
 
-# Restrict to specific emails
+# Restrict to specific emails (they log in via Hypha)
 svamp serve dashboard ./outputs --access alice@x.com,bob@y.com
+
+# Listed publicly (anyone, no URL secrecy assumed)
+svamp serve dashboard ./outputs --public
 ```
 
+⚠️ **Do not assume the default is login-gated.** It is not — `svamp serve` prints the tier back to
+you as `link (capability URL — anyone with the URL can view)`. Before serving anything internal,
+either pass `--access owner` or accept that possession of the URL is possession of the content.
+Because the hostname is the secret, there is no way to revoke access to a `link` mount other than
+`--regenerate-url`, which mints a new one; deleting the mount does not un-publish what was already
+fetched.
+
+Use the URL exactly as `serve`/`serve list` prints it — never compose it by hand, and never append a
+path to it.
+
 Post the resulting URL as a normal markdown link. The svamp client doesn't manage shares — `svamp serve` is the dedicated tool.
 
 ## Rendering environment

package/bin/skills/crew/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: crew
-version: 0.3.2
+version: 0.3.4
 description: Coordinate work as a lead delegating features to managed worktree children — spawn a child per feature on its own branch, run it as a loop until done, and pull finished work back to the base branch with merge-and-close. Use when one session should run a project and farm out features instead of piling many sessions on one working tree.
 ---
 
@@ -13,10 +13,27 @@ the same working tree.
 
 A **crew** is emergent, not a heavy object: it is just plain sessions linked by
 `parentSessionId` + a `crew.role`. A **feature child** always (a) has the lead as its
-parent, (b) lives in its own `.dev/worktree/<n>` on its own branch, and (c) runs as a loop
-with `parent: <lead>` — i.e. **the lead is the judge** that decides when the child is done
-(this is the `parent`-routed case of the loop … until gate; the loop engine is Claude Code's
-native `/goal`, driven from `svamp session loop`).
+parent and (b) lives in its own `.dev/worktree/<n>` on its own branch.
+
+> **⚠ Whether the child is GATED depends on how you started it (#1604).**
+>
+> | You ran | Gate |
+> |---|---|
+> | `crew start "<brief>"` (no `--oracle`) | **none** — the child works, then runs `feature done` when *it* decides it is complete. Nothing judges it. |
+> | `crew start "<brief>" --oracle "<cmd>"` | a loop with `parent: <lead>` — the oracle gates each stop and **the lead is the judge** of done. |
+>
+> The CLI tells you which you got: it prints `gate: none — child runs 'feature done' when complete`
+> on the ungated path. This is deliberate, not an oversight — `featureStart` attaches the loop only
+> when there is an oracle, because a no-oracle child would otherwise fire `approved` on its very
+> first stop, which is worse than no gate at all.
+>
+> **What this means for you as a lead:** if the feature has any objective success condition, pass
+> `--oracle`. Without one you are the only thing standing between a child and a premature
+> `feature done`, so read the merge request rather than trusting the verdict.
+>
+> This block used to claim every feature child runs as a loop with the lead as judge. It did not,
+> and an agent that believed it would skip `--oracle`, skip checking on the child, and read "no
+> verdict yet" as "still working" rather than "nothing will ever judge this".
 
 ## The verbs (these are your tools)
 
@@ -29,10 +46,17 @@ svamp feature start "<brief>" [--oracle "<test cmd>"] [--base <branch>] [--actio
         #   nudge (default) = normal urgency, the lead sees it on its next idle turn;
         #   block = urgent, which WAKES the lead. Use block for a child whose result
         #   gates your next step; nudge for one you will collect when convenient.
-        # --issue <id> links the child to an existing backlog issue: its verdict routes
-        #   ONTO that issue. This is the seam that joins crew to the backlog the loop
-        #   oracle (`svamp issue pending`) reads — without it the child's result is
-        #   invisible to the loop, so pass it whenever the feature came from an issue.
+        # --issue <id> links the child to an EXISTING backlog issue: its verdict routes
+        #   ONTO that issue instead of onto a new one. This is the seam that joins crew
+        #   to the backlog the loop oracle (`svamp issue pending`) reads.
+        #   Pass it whenever the feature came from an issue, so the verdict lands on the
+        #   item you are already tracking rather than creating a second one to reconcile.
+        #   NOT passing it does NOT make the child invisible: `crew start` auto-creates a
+        #   tracking issue at status `in_progress`, and `isPendingIssue` counts anything
+        #   that is neither `archived` nor `paused` — so an unlinked child still holds the
+        #   loop oracle RED until its issue is resolved. Do not treat an unlinked child as
+        #   fire-and-forget; a lead waiting for `svamp issue pending` to empty will wait
+        #   for it.
         # --max N bounds the child's loop iterations (a hard stop, unlike the soft
         #   progress-based cap); --model <id> picks the child's model.
 svamp feature start "<brief>" --no-worktree --dir <folder>
@@ -81,8 +105,15 @@ You coordinate the project. Keep your own working tree (the base branch, e.g. `m
 
 ## If you are a FEATURE CHILD
 
-You own one feature in an isolated worktree on your own branch, reporting to your lead. Your
-brief (and your loop's "until" criteria) is injected each turn.
+You own one feature in an isolated worktree on your own branch, reporting to your lead.
+
+**Your brief arrives ONCE, as the first message of this session — it is not repeated (#1607).**
+Nothing re-injects it, so keeping it in view is your job: re-read it before you decide you are done,
+and restate it in your own words in `feature report` notes so drift is visible to your lead. (If your
+lead started you *with* `--oracle`, the native `/goal` gate does restate the "until" criteria at each
+turn-end — but on the default no-oracle path there is no loop at all. See the gate table above.)
+This line previously promised per-turn injection of both; no such mechanism was ever built, and
+`docs/crew-design.md` §7 already recorded those hooks as never shipped.
 
 - **Work in your worktree.** Implement the feature. Commit as you go. Keep the oracle/tests green.
 - **Report.** Use `svamp feature report "<note>"` for milestones and `--blocker` when stuck —

package/dist/{acpBackend-BZ1MJhQV.mjs → acpBackend-UaBD4zL3.mjs}

@@ -558,6 +558,9 @@ class AcpBackend {
               return { outcome: { outcome: "selected", optionId: "cancel" } };
             }
           }
+          if (!this.options.permissionHandler) {
+            return { outcome: { outcome: "cancelled" } };
+          }
           const proceedOnceOption = options.find(
             (opt) => opt.optionId === "proceed_once" || typeof opt.name === "string" && opt.name.toLowerCase().includes("once")
           );

package/dist/{agentCommands-B2Nl7Fof.mjs → agentCommands-D91RcjqQ.mjs}

@@ -79,7 +79,7 @@ async function sessionSetTitle(title) {
 }
 async function sessionSetProjectDescription(description) {
   const dir = process.cwd();
-  const { projectName, writeProjectInfo, sanitizeDescription, projectInfoPath } = await import('./run-DkjwF-zA.mjs').then(function (n) { return n.p; });
+  const { projectName, writeProjectInfo, sanitizeDescription, projectInfoPath } = await import('./run-DDmeV9K3.mjs').then(function (n) { return n.p; });
   const desc = sanitizeDescription(description, 240);
   if (!desc) {
     console.error("Project description is empty.");
@@ -382,7 +382,7 @@ async function sessionBroadcast(action, args) {
   console.log(`Broadcast sent: ${action}`);
 }
 async function connectToMachineService() {
-  const { connectAndGetMachine } = await import('./commands-D1pZq3d4.mjs');
+  const { connectAndGetMachine } = await import('./commands-B-yczgz_.mjs');
   return connectAndGetMachine();
 }
 function buildInboxMessage(args) {
@@ -465,7 +465,7 @@ async function inboxSend(targetSessionId, opts) {
     console.error("Message body is required.");
     process.exit(1);
   }
-  const { connectAndResolveSession } = await import('./commands-D1pZq3d4.mjs');
+  const { connectAndResolveSession } = await import('./commands-B-yczgz_.mjs');
   let server;
   try {
     const { targetId, messageId } = await inboxSendCore(
@@ -520,7 +520,7 @@ async function inboxReply(messageId, body, opts) {
     console.error("SVAMP_SESSION_ID not set. This command must be run inside a Svamp session.");
     process.exit(1);
   }
-  const { connectAndResolveSession } = await import('./commands-D1pZq3d4.mjs');
+  const { connectAndResolveSession } = await import('./commands-B-yczgz_.mjs');
   const { server: localServer, machine: localMachine } = await connectToMachineService();
   let localDisconnected = false;
   const disconnectLocal = async () => {

package/dist/{claudeAuth-iAerdaMr.mjs → claudeAuth-C0wc1GTM.mjs}

@@ -182,8 +182,52 @@ function applyPromptCacheTtl(spawnEnv) {
   if (explicit1h === void 0) spawnEnv.ENABLE_PROMPT_CACHING_1H = "1";
 }
 const BETA_1M_TOKEN = "context-1m-2025-08-07";
+let longContext1MEntitlement = "unknown";
+function setLongContext1MEntitlement(state) {
+  longContext1MEntitlement = state;
+}
+function getLongContext1MEntitlement() {
+  return longContext1MEntitlement;
+}
+function isLongContextBetaUnavailableError(err) {
+  const text = typeof err === "string" ? err : err?.message ?? String(err ?? "");
+  return /long context beta is not (yet )?available/i.test(text);
+}
+const LONG_CONTEXT_ENTITLEMENT_RECHECK_MS = Number(process.env.SVAMP_1M_ENTITLEMENT_RECHECK_MS || 60 * 60 * 1e3);
+async function probeLongContext1MEntitlement(baseUrl, apiKey, fetchImpl = fetch) {
+  if (!baseUrl || !apiKey) return longContext1MEntitlement;
+  try {
+    const resp = await fetchImpl(`${baseUrl.replace(/\/$/, "")}/v1/messages`, {
+      method: "POST",
+      headers: {
+        "content-type": "application/json",
+        authorization: `Bearer ${apiKey}`,
+        "anthropic-beta": BETA_1M_TOKEN
+      },
+      body: JSON.stringify({
+        model: "claude-haiku-4-5-20251001",
+        max_tokens: 1,
+        messages: [{ role: "user", content: "x" }]
+      }),
+      signal: AbortSignal.timeout(2e4)
+    });
+    if (resp.ok) {
+      setLongContext1MEntitlement("entitled");
+      return "entitled";
+    }
+    const text = await resp.text().catch(() => "");
+    if (isLongContextBetaUnavailableError(text)) {
+      setLongContext1MEntitlement("not-entitled");
+      return "not-entitled";
+    }
+    return longContext1MEntitlement;
+  } catch {
+    return longContext1MEntitlement;
+  }
+}
 function applyLongContext1M(spawnEnv) {
   if (envFlagEnabled(spawnEnv.CLAUDE_CODE_DISABLE_1M_CONTEXT) || envFlagEnabled(process.env.CLAUDE_CODE_DISABLE_1M_CONTEXT)) return;
+  if (longContext1MEntitlement !== "entitled") return;
   const existing = spawnEnv.ANTHROPIC_BETAS ?? process.env.ANTHROPIC_BETAS ?? "";
   const betas = existing.split(",").map((b) => b.trim()).filter(Boolean);
   if (!betas.includes(BETA_1M_TOKEN)) betas.push(BETA_1M_TOKEN);
@@ -244,4 +288,4 @@ function applyClaudeProxyEnv(spawnEnv) {
 }
 const MANAGED_ENV_KEYS = Array.from(MANAGED_KEYS);
 
-export { BETA_1M_TOKEN, EXAMPLE_HYPHA_PROXY_URL, MANAGED_ENV_KEYS, applyClaudeProxyEnv, getClaudeAuthStatus, normalizeProxyUrl, resolveHyphaProxyUrl, setClaudeAuthCustom, setClaudeAuthHyphaProxy, setClaudeAuthLogin, updateEnvFile };
+export { BETA_1M_TOKEN, EXAMPLE_HYPHA_PROXY_URL, LONG_CONTEXT_ENTITLEMENT_RECHECK_MS, MANAGED_ENV_KEYS, applyClaudeProxyEnv, getClaudeAuthStatus, getLongContext1MEntitlement, isLongContextBetaUnavailableError, normalizeProxyUrl, probeLongContext1MEntitlement, resolveHyphaProxyUrl, setClaudeAuthCustom, setClaudeAuthHyphaProxy, setClaudeAuthLogin, setLongContext1MEntitlement, updateEnvFile };