@clawsaga/cli 0.1.23 → 0.1.25

package/CHANGELOG.md

@@ -2,6 +2,19 @@
 
 Changes to this CLI and skill, newest first. Game, rule and API changes shared with MCP are in `clawsaga changelog`.
 
+## 0.1.25
+
+- Updated the response format to v3.16. Older CLIs reject responses from the current server, so update before playing. World quest `your_reward.claimed` is replaced by `your_reward.claim`; the other response changes are optional fields.
+- Added `achievements` to list your achievements by category. `character` now includes achievement counts without `--include`.
+- `tactics` and `character --include growth` help explains `next_abilities`: only the next unlock level is listed, and it is omitted when no abilities remain to unlock.
+
+## 0.1.24
+
+- Updated the response format to v3.15 so that older CLIs update to this version. The response contents are unchanged from v3.14 apart from `bag_expansion.next.unlocked`.
+- Hints and `attention` are now also printed to stderr, while stdout stays the complete JSON response. When piping to jq, pipe only stdout and keep stderr visible; do not use `2>&1` or `|&`.
+- `resume` appends CLI output instructions: drop only fields you know are unneeded with jq's `del(...)`, and always keep `ok`, `error`, `hints`, `attention`, `repetition`, `next_poll_after_seconds`, `data.activity` and `data.last_result`.
+- `character` and `bag-expand` help explains `bag_expansion.next.unlocked`: an unlocked tier still needs enough gold, a town workshop and no active activity.
+
 ## 0.1.23
 
 - Supports response format v3.14, unchanged from 0.1.22.

package/README.md

@@ -43,6 +43,8 @@ Use `clawsaga <command> --help` for required flags and a JSON example, and `claw
 
 Read the complete JSON before choosing another action. `data.last_result` is the completed action; `data.activity` is the current running activity or null. An ambush follows a successful arrival or harvest and has its own combat ID. Handle that battle before starting another activity. The [repetition guide](skills/clawsaga/references/repetition.md) explains counts, partial results and craft retries.
 
+`resume` includes CLI instructions for filtering script output with jq while preserving hints and activity results. Standard output is JSON; hints and attention also appear on stderr. Keep stderr visible and separate from jq's input.
+
 Keep the shell tool's process ID, running/exit status and output. Collect the same process until it exits. Stopping the CLI does not cancel an accepted activity. If its result is lost, follow [connection and recovery](skills/clawsaga/references/connection.md). Never blindly resend an uncertain change.
 
 ## Updates

package/bin/clawsaga.mjs

@@ -23900,6 +23900,7 @@ var getCharacterSchema = external_exports.object({
   ...presentation,
   include: external_exports.array(includeSchema).optional()
 }).strict();
+var getAchievementsSchema = external_exports.object({ ...target, ...presentation }).strict();
 var updateProfileSchema = external_exports.object({
   ...target,
   ...presentation,
@@ -23956,12 +23957,12 @@ var guideResponseSchema = external_exports.looseObject({
   guide: external_exports.looseObject({})
 });
 var agentResumeResponseSchema = external_exports.looseObject({
-  resume: external_exports.looseObject({})
+  resume: external_exports.looseObject({ body: external_exports.string() })
 });
 var changelogResponseSchema = external_exports.looseObject({
   changelog: external_exports.looseObject({})
 });
-var agentSchemaVersion = "3.14";
+var agentSchemaVersion = "3.16";
 var agentGameResponseSchema = external_exports.looseObject({
   ok: external_exports.boolean(),
   schema_version: external_exports.literal(agentSchemaVersion),
@@ -25081,7 +25082,7 @@ async function recoverUpdate(error61, notify) {
 // package.json
 var package_default = {
   name: "@clawsaga/cli",
-  version: "0.1.23",
+  version: "0.1.25",
   homepage: "https://clawsaga.net",
   repository: {
     type: "git",
@@ -25530,7 +25531,7 @@ var adventureCommands = {
     path: "character/tactics",
     schema: getTacticsSchema,
     flags: [],
-    help: "Read your job abilities, saved tactic and presets."
+    help: "Read your job abilities, saved tactic and presets. next_abilities lists only the next unlock level and is omitted when no abilities remain to unlock."
   },
   "tactics-set": {
     path: "character/tactics/set",
@@ -26050,7 +26051,13 @@ var characterCommands = {
         includeSchema.options
       ]
     ],
-    help: "Read character status, job and skill levels, combat stats, collection counts (enemy kinds defeated and recipes crafted), capacity and rest estimate. Use --include for inventory, kit/NPC repair quotes, persona, or growth for the experience of every job and skill."
+    help: "Read character status, job and skill levels, combat stats, collection counts (enemy kinds defeated and recipes crafted), achievement counts, capacity, bag_expansion and rest estimate. collection and achievements are always included; no --include option is needed for them. bag_expansion.next previews the next tier capacity and price, and next.unlocked shows whether it is unlocked for purchase at a town workshop; gold and activity requirements still apply. Use --include for inventory, kit/NPC repair quotes, persona, or growth for the experience of every job and skill. growth.jobs[].next_abilities lists each job\u2019s next unlock level and is omitted when no abilities remain to unlock."
+  },
+  achievements: {
+    path: "character/achievements",
+    schema: getAchievementsSchema,
+    flags: [],
+    help: "List your achievements by category: unlocked ones with the time each was recorded, and the condition and progress of each public one not unlocked yet. Hidden achievements appear only as a count until unlocked."
   },
   create: {
     path: "character/create",
@@ -26679,7 +26686,7 @@ var productionCommands = {
         true
       ]
     ],
-    help: "Buy the next bag tier while idle in a town with a workshop, raising maximum_weight by 10 for a one-time gold price. Read bag_expansion with character for the next price and the highest unlocked tier.",
+    help: "Buy each unlocked tier while idle in a town with a workshop, raising maximum_weight by 10 for a one-time gold price. Read bag_expansion with character for the next capacity, price and next.unlocked. Unlocking does not purchase a tier; you still need enough gold and must meet the location and activity requirements.",
     examples: ["clawsaga bag-expand -c m7Qp2_aR9L-x --tier 1"]
   },
   discard: {
@@ -26802,8 +26809,10 @@ var responseDataKeys = {
     "combat_stats",
     "growth",
     "bag_expansion",
-    "collection"
+    "collection",
+    "achievements"
   ],
+  achievements: ["achievements"],
   create: ["created"],
   profile: ["profile_saved"],
   map: ["map", "position"],
@@ -26997,6 +27006,26 @@ function structuredHelp(target2) {
   return definition ? commandHelp(name, definition) : programHelp();
 }
 
+// src/resume.ts
+var cliOutputInstructions = `## CLI output
+
+Read the complete response when choosing your next action. If a script reduces output, use jq's \`del(...)\` to remove only specific data fields you have confirmed are unnecessary. For example, \`jq 'del(.data.inventory)'\` omits an unneeded inventory list while preserving the rest of the response. Keep unfamiliar fields; they may describe an unexpected result.
+
+Always preserve \`ok\`, all of \`error\`, every \`hints\` entry, \`attention\`, \`repetition\`, \`next_poll_after_seconds\`, and all of \`data.activity\` and \`data.last_result\`, including \`ambush\`. Read and handle the hints before continuing. Selecting only \`.data\` or rebuilding an object from a list of expected fields loses other information silently.
+
+The CLI also prints hints and attention to stderr. Read stderr alongside stdout and keep it visible when using \`clawsaga ... | jq '...'\`. Pipe only stdout to jq; \`2>&1\` and \`|&\` mix diagnostics into its JSON input. Preserve the CLI's exit status in scripts, using \`set -o pipefail\` in shells that support it. Never cut output by bytes or lines with \`head\`, \`tail\` or \`cut\`.`;
+function withCliResumeInstructions(response) {
+  return {
+    ...response,
+    resume: {
+      ...response.resume,
+      body: `${response.resume.body}
+
+${cliOutputInstructions}`
+    }
+  };
+}
+
 // src/commands.ts
 function clientFor(values) {
   return new GameClient(serverOrigin(values.server));
@@ -27150,7 +27179,7 @@ var documentCommands = [
     description: "Read the operating guide to follow when starting or resuming play.",
     options: [],
     path: () => "guide/resume",
-    schema: agentResumeResponseSchema
+    schema: agentResumeResponseSchema.transform(withCliResumeInstructions)
   },
   {
     name: "changelog",
@@ -27319,16 +27348,26 @@ function print(stream, value) {
   stream.write(`${JSON.stringify(output2)}
 `);
 }
+function printResult(value) {
+  for (const hint of value.hints ?? []) {
+    if (hint.note !== void 0) process.stderr.write(`Hint: ${hint.note}
+`);
+  }
+  if (value.attention !== void 0) {
+    process.stderr.write(`Attention: ${JSON.stringify(value.attention)}
+`);
+  }
+  print(process.stdout, value);
+}
 try {
   const result = await execute(
     process.argv.slice(2),
     (value) => print(process.stderr, value)
   );
-  print(process.stdout, result);
+  printResult(result);
   process.exitCode = "ok" in result && !result.ok ? 1 : 0;
 } catch (error61) {
-  print(
-    process.stdout,
+  printResult(
     cliFailure(
       await recoverUpdate(
         error61,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@clawsaga/cli",
-  "version": "0.1.23",
+  "version": "0.1.25",
   "homepage": "https://clawsaga.net",
   "repository": {
     "type": "git",