aitasks 1.4.5 → 1.4.6

package/README.md

@@ -67,6 +67,21 @@ export AITASKS_JSON=true
 
 ---
 
+## Global Options
+
+| Option | Description |
+|---|---|
+| `-C, --dir <path>` | Run as if `aitasks` were started in `<path>` instead of the current directory |
+| `--json` | Output machine-readable JSON (available on most commands) |
+| `--version` | Print the installed version |
+| `--help` | Show help for any command, e.g. `aitasks update --help` |
+
+```sh
+aitasks -C /path/to/project board
+```
+
+---
+
 ## Commands
 
 ### Setup
@@ -75,6 +90,7 @@ export AITASKS_JSON=true
 |---|---|
 | `aitasks init` | Initialize a task database in the current project |
 | `aitasks init --with-review` | Initialize with review enforcement (agents cannot mark done without a passing review) |
+| `aitasks init --update` | Refresh the agent instructions block in an existing CLAUDE.md / AGENTS.md / GEMINI.md to the latest version |
 | `aitasks onboard` | Print or inject agent protocol instructions into CLAUDE.md / AGENTS.md |
 
 ### Task Discovery
@@ -82,7 +98,7 @@ export AITASKS_JSON=true
 | Command | Description |
 |---|---|
 | `aitasks list` | List all tasks, sorted by priority |
-| `aitasks list --status ready` | Filter by status (`ready`, `in_progress`, `blocked`, `review`, `done`) |
+| `aitasks list --status ready` | Filter by status (`backlog`, `ready`, `in_progress`, `blocked`, `review`, `done`) |
 | `aitasks next` | Show the highest-priority unblocked ready task |
 | `aitasks next --claim --agent <id>` | Auto-claim and start the best task |
 | `aitasks show <id>` | Full detail on a specific task (includes time tracking) |
@@ -99,6 +115,7 @@ export AITASKS_JSON=true
 | `aitasks start <id...> --agent <id>` | Begin active work on task(s) |
 | `aitasks note <id> <text>` | Add an implementation note |
 | `aitasks check <id> <n> --evidence <text>` | Verify acceptance criterion n |
+| `aitasks update <id>` | Update task fields (title, description, priority, type, status, acceptance criteria) — see [`update` Flags](#update-flags) |
 | `aitasks done <id...> --agent <id>` | Mark task(s) complete (all criteria must be verified; must be in `review` status if enforcement is on) |
 | `aitasks review <id...> --agent <id>` | Submit task(s) for review (moves to `review` status) |
 | `aitasks reject <id> --reason <text>` | Reject a task in review, send it back to `in_progress` with feedback |
@@ -148,6 +165,41 @@ aitasks create \
 
 ---
 
+## `update` Flags
+
+Change any field on an existing task:
+
+```sh
+aitasks update TASK-001 \
+  --title "New title" \         # Replace the title
+  --desc "New description" \    # Replace the description
+  --priority high \             # critical | high | medium | low
+  --type bug \                  # feature | bug | chore | spike
+  --status ready                # Manually override status (use with care)
+```
+
+### Editing acceptance criteria
+
+Criterion indices are **0-based**, matching what `aitasks show` and `aitasks check` display. Use **only one** of these flags per invocation:
+
+| Flag | Effect |
+|---|---|
+| `--ac <text>` | **Append** a new criterion (repeatable). Does *not* replace existing ones — re-passing an existing criterion duplicates it. |
+| `--set-ac <index>=<text>` | Replace just the criterion at `<index>` in place. |
+| `--remove-ac <index>` | Delete the criterion at `<index>`. |
+| `--replace-ac <list>` | Overwrite the **entire** list (newline-separated). |
+
+```sh
+aitasks update TASK-001 --ac "New criterion to append"
+aitasks update TASK-001 --set-ac 1="Returns 404 with a JSON error body"
+aitasks update TASK-001 --remove-ac 2
+aitasks update TASK-001 --replace-ac $'First criterion\nSecond criterion\nThird criterion'
+```
+
+> **Re-verify after editing.** `--set-ac` clears any prior verification of that criterion (since the wording changed), and `--remove-ac` re-indexes the remaining verifications. Run `aitasks check` again for affected criteria before marking the task done.
+
+---
+
 ## Agent Protocol
 
 When you run `aitasks init`, it automatically injects a full agent protocol into `CLAUDE.md`, `AGENTS.md`, or `GEMINI.md` (whichever exists, or creates `AGENTS.md`). This tells the AI agent exactly how to use `aitasks`.
@@ -163,6 +215,14 @@ aitasks onboard --json        # output as JSON string
 
 The injected instructions automatically adapt to the project's review enforcement setting — if `--with-review` is enabled, agents receive the full review workflow (with `aitasks review`, sub-agent approval, and `aitasks reject`) instead of the standard completion flow.
 
+**Updating after an upgrade:** Re-running `aitasks init` on a project whose agent file already contains instructions leaves them untouched (and prints a hint). To pull the latest protocol after upgrading `aitasks`, run:
+
+```sh
+aitasks init --update
+```
+
+This replaces only the block between the `<!-- aitasks:instructions -->` markers in place — the rest of your `CLAUDE.md` / `AGENTS.md` is preserved.
+
 ---
 
 ## Review Enforcement

package/dist/index.js

@@ -1890,7 +1890,7 @@ var require_commander = __commonJS((exports) => {
 var require_package = __commonJS((exports, module) => {
   module.exports = {
     name: "aitasks",
-    version: "1.4.5",
+    version: "1.4.6",
     description: "CLI task management tool built for AI agents",
     type: "module",
     bin: {
@@ -29402,6 +29402,18 @@ Creating subtasks:
 aitasks create --title "Write unit tests for auth" --desc "Add unit tests covering all auth edge cases" --ac "All tests pass" --ac "Coverage \u2265 90%" --parent TASK-001 --priority high --type chore --agent $AITASKS_AGENT_ID
 \`\`\`
 
+Editing acceptance criteria (if the requirements change or were worded wrong):
+\`\`\`bash
+aitasks update TASK-001 --ac "A brand-new criterion to append"       # add one
+aitasks update TASK-001 --set-ac 1="Returns 404 with a JSON error body"  # fix criterion #1 in place
+aitasks update TASK-001 --remove-ac 2                                # delete criterion #2
+aitasks update TASK-001 --replace-ac $'first\\nsecond\\nthird'        # replace the whole list
+\`\`\`
+Indices are 0-based (matching \`aitasks show\` and \`aitasks check\`). Use exactly one of these
+flags per call. \`--ac\` only APPENDS \u2014 re-passing existing criteria duplicates them; use
+\`--set-ac\` to correct one criterion or \`--replace-ac\` to rewrite all of them. After
+\`--set-ac\`, re-run \`aitasks check\` for that index since editing the wording clears its prior verification.
+
 If you discover your task is blocked by something:
 \`\`\`bash
 aitasks block TASK-001 --on TASK-002,TASK-003
@@ -29540,7 +29552,11 @@ aitasks show <id>                           Full task detail (includes time trac
 aitasks search <query>                      Search titles, descriptions, notes
 aitasks deps <id>                           Show dependency tree
 aitasks create --title <t> --desc <d> --ac <c> [--ac <c> ...] --agent <id>   Create a task
-aitasks update <id> [--status|--priority|--title|--desc|--ac|--type]   Update task fields
+aitasks update <id> [--status|--priority|--title|--desc|--type]        Update task fields
+aitasks update <id> --ac <text>             Append a new acceptance criterion
+aitasks update <id> --set-ac <n>=<text>     Replace a single criterion at index n (0-based)
+aitasks update <id> --remove-ac <n>         Remove a single criterion at index n (0-based)
+aitasks update <id> --replace-ac <list>     Replace ALL criteria (newline-separated)
 aitasks claim <id...> --agent <id>          Claim task(s) - supports patterns like TASK-0*
 aitasks start <id...> --agent <id>          Begin work on task(s)
 aitasks note <id> <text> --agent <id>       Add implementation note
@@ -29587,7 +29603,7 @@ function injectOrCreateAgentFile(projectRoot, version, reviewRequired = false, f
         return { filePath: existing, action: "skipped" };
       const replaced = replaceInstructionsBlock(content, instructions);
       writeFileSync2(existing, replaced, "utf8");
-      return { filePath: existing, action: "appended" };
+      return { filePath: existing, action: "updated" };
     }
     const separator = content.endsWith(`
 `) ? `
@@ -29678,7 +29694,7 @@ function getVersion() {
 }
 
 // src/commands/init.ts
-var initCommand = new Command("init").description("Initialize AITasks in the current project").option("--skip-agent-file", "Skip injecting agent instructions into CLAUDE.md/AGENTS.md/GEMINI.md").option("--with-review", "Enforce review gate: agents cannot mark tasks done without a passing review").addHelpText("after", `
+var initCommand = new Command("init").description("Initialize AITasks in the current project").option("--skip-agent-file", "Skip injecting agent instructions into CLAUDE.md/AGENTS.md/GEMINI.md").option("--with-review", "Enforce review gate: agents cannot mark tasks done without a passing review").option("--update", "Refresh the agent instructions block in an existing CLAUDE.md/AGENTS.md/GEMINI.md to the latest version").addHelpText("after", `
 Examples:
   $ aitasks init
       Initialize in current directory, inject agent instructions into CLAUDE.md/AGENTS.md.
@@ -29689,6 +29705,11 @@ Examples:
       marked done. Running this on an existing project also updates the agent
       instructions file with the review workflow.
 
+  $ aitasks init --update
+      Re-inject the latest agent instructions, replacing the existing block
+      between the AITasks markers in place. The rest of the file is left
+      untouched. Use after upgrading aitasks to pull new protocol docs.
+
   $ aitasks init --skip-agent-file
       Initialize without touching any CLAUDE.md / AGENTS.md file.`).action(async (opts) => {
   const root = findProjectRoot();
@@ -29701,7 +29722,8 @@ Examples:
       console.log(source_default.green("  \u2713") + "  Review enforcement enabled.");
     }
     if (!opts.skipAgentFile) {
-      const result = injectOrCreateAgentFile(root, getVersion(), !!opts.withReview, !!opts.withReview);
+      const force = !!opts.withReview || !!opts.update;
+      const result = injectOrCreateAgentFile(root, getVersion(), !!opts.withReview, force);
       printAgentFileResult(result);
     }
     return;
@@ -29722,7 +29744,7 @@ Examples:
   }
   console.log("");
   if (!opts.skipAgentFile) {
-    const result = injectOrCreateAgentFile(root, getVersion(), !!opts.withReview);
+    const result = injectOrCreateAgentFile(root, getVersion(), !!opts.withReview, !!opts.update);
     printAgentFileResult(result);
   }
   console.log(source_default.dim("  Run `aitasks create` to add your first task."));
@@ -29737,8 +29759,12 @@ function printAgentFileResult(result) {
     case "appended":
       console.log(source_default.green("  \u2713") + `  Appended agent instructions to ${source_default.bold(rel)}`);
       break;
+    case "updated":
+      console.log(source_default.green("  \u2713") + `  Updated agent instructions in ${source_default.bold(rel)}`);
+      break;
     case "skipped":
       console.log(source_default.dim(`  \u2500  ${rel} already contains AITasks instructions`));
+      console.log(source_default.dim(`     Run \`aitasks init --update\` to refresh them to the latest version.`));
       break;
   }
   console.log("");
@@ -41672,7 +41698,20 @@ var nextCommand = new Command("next").description("Show the next best task to wo
 });
 
 // src/commands/update.ts
-var updateCommand = new Command("update").description("Update task fields").argument("<taskId>", "Task ID").option("-t, --title <title>", "New title").option("-d, --desc <description>", "New description").option("-a, --ac <criterion>", "Add an acceptance criterion (repeatable)", collect2, []).option("--replace-ac <criteria>", "Replace ALL acceptance criteria (newline-separated)").option("-p, --priority <priority>", "New priority: critical|high|medium|low").option("--type <type>", "New type: feature|bug|chore|spike").option("--status <status>", "Manually set status (use with care)").option("--json", "Output as JSON").action((taskId, opts) => {
+var updateCommand = new Command("update").description("Update task fields").argument("<taskId>", "Task ID").option("-t, --title <title>", "New title").option("-d, --desc <description>", "New description").option("-a, --ac <criterion>", "Add an acceptance criterion (repeatable)", collect2, []).option("--replace-ac <criteria>", "Replace ALL acceptance criteria (newline-separated)").option("--set-ac <index=text>", 'Replace a single acceptance criterion by 0-based index, e.g. --set-ac 1="New text"').option("--remove-ac <index>", "Remove a single acceptance criterion by 0-based index").option("-p, --priority <priority>", "New priority: critical|high|medium|low").option("--type <type>", "New type: feature|bug|chore|spike").option("--status <status>", "Manually set status (use with care)").option("--json", "Output as JSON").addHelpText("after", `
+Acceptance criteria editing:
+  --ac <text>            Append a new criterion (repeatable). Does NOT replace existing ones.
+  --set-ac <index=text>  Replace just the criterion at <index> (0-based). Re-verify it afterward \u2014
+                         any prior verification of that criterion is cleared since the wording changed.
+  --remove-ac <index>    Delete the criterion at <index> (0-based). Verifications are re-indexed.
+  --replace-ac <list>    Overwrite the entire list (newline-separated).
+
+  Only one acceptance-criteria mode may be used per invocation.
+
+Examples:
+  $ aitasks update TASK-001 --set-ac 1="Returns 404 with a JSON error body"
+  $ aitasks update TASK-001 --remove-ac 2
+  $ aitasks update TASK-001 --ac "New criterion to append"`).action((taskId, opts) => {
   requireInitialized();
   const json = isJsonMode(opts.json);
   const id = taskId.toUpperCase();
@@ -41690,11 +41729,39 @@ var updateCommand = new Command("update").description("Update task fields").argu
     changes.type = opts.type;
   if (opts.status)
     changes.status = opts.status;
-  if (opts.replaceAc) {
+  const acModes = [
+    opts.ac.length > 0,
+    opts.replaceAc !== undefined,
+    opts.setAc !== undefined,
+    opts.removeAc !== undefined
+  ].filter(Boolean).length;
+  if (acModes > 1) {
+    exitError("Use only one of --ac, --replace-ac, --set-ac, or --remove-ac per invocation.", json);
+  }
+  if (opts.replaceAc !== undefined) {
     changes.acceptance_criteria = opts.replaceAc.split(`
 `).map((s) => s.trim()).filter(Boolean);
   } else if (opts.ac.length > 0) {
     changes.acceptance_criteria = [...task.acceptance_criteria, ...opts.ac];
+  } else if (opts.setAc !== undefined) {
+    const eq = opts.setAc.indexOf("=");
+    if (eq < 0) {
+      exitError('--set-ac requires the format <index>=<text>, e.g. --set-ac 1="New criterion".', json);
+    }
+    const index = parseCriterionIndex(opts.setAc.slice(0, eq), task.acceptance_criteria.length, json);
+    const text = opts.setAc.slice(eq + 1).trim();
+    if (!text)
+      exitError("--set-ac text cannot be empty.", json);
+    const criteria = [...task.acceptance_criteria];
+    criteria[index] = text;
+    changes.acceptance_criteria = criteria;
+    if (task.test_results.some((r2) => r2.index === index)) {
+      changes.test_results = task.test_results.filter((r2) => r2.index !== index);
+    }
+  } else if (opts.removeAc !== undefined) {
+    const index = parseCriterionIndex(opts.removeAc, task.acceptance_criteria.length, json);
+    changes.acceptance_criteria = task.acceptance_criteria.filter((_2, i) => i !== index);
+    changes.test_results = task.test_results.filter((r2) => r2.index !== index).map((r2) => r2.index > index ? { ...r2, index: r2.index - 1 } : r2);
   }
   if (Object.keys(changes).length === 0) {
     exitError("No changes specified. Use --help to see available options.", json);
@@ -41716,6 +41783,16 @@ function collect2(val, acc) {
   acc.push(val);
   return acc;
 }
+function parseCriterionIndex(raw, total, json) {
+  const index = parseInt(raw.trim(), 10);
+  if (isNaN(index) || index < 0) {
+    exitError(`Invalid criterion index "${raw.trim()}". Use a non-negative integer: 0, 1, 2, \u2026`, json);
+  }
+  if (index >= total) {
+    exitError(`Index ${index} out of range \u2014 task has ${total} criteria (0-based, so valid range is 0\u2013${total - 1}).`, json);
+  }
+  return index;
+}
 
 // src/display/board.tsx
 var import_react29 = __toESM(require_react(), 1);
@@ -43533,4 +43610,4 @@ program2.parseAsync(process.argv).catch((err) => {
   process.exit(1);
 });
 
-//# debugId=DD10AE067D1A79CC64756E2164756E21
+//# debugId=49CA369B086466F464756E2164756E21

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "aitasks",
-  "version": "1.4.5",
+  "version": "1.4.6",
   "description": "CLI task management tool built for AI agents",
   "type": "module",
   "bin": {