agentpack-cli 0.2.0__tar.gz → 0.2.1__tar.gz

{agentpack_cli-0.2.0 → agentpack_cli-0.2.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agentpack-cli
-Version: 0.2.0
+Version: 0.2.1
 Summary: Task-aware context packing for AI coding agents — Claude, Cursor, Windsurf, Codex, and Antigravity
 License: MIT
 License-File: LICENSE
@@ -44,7 +44,7 @@ Description-Content-Type: text/markdown
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 [![CI](https://github.com/vishal2612200/agentpack/actions/workflows/ci.yml/badge.svg)](https://github.com/vishal2612200/agentpack/actions/workflows/ci.yml)
 
-> **Status: alpha (v0.2.0).** Works, tested, used in real sessions. Python and JavaScript/TypeScript are the best-supported languages. Not yet validated across a wide range of repos. API may change before 1.0.
+> **Status: alpha (v0.2.1).** Works, tested, used in real sessions. Python and JavaScript/TypeScript are the best-supported languages. Not yet validated across a wide range of repos. API may change before 1.0.
 >
 > **Platform note:** macOS and Linux are fully supported. Windows support is not yet implemented (git hooks use POSIX shell; the Claude Code session hooks use `python3`/`rm -f`). Contributions welcome.
 
@@ -100,7 +100,8 @@ The npm package is a Node launcher around the Python implementation. It installs
 ```bash
 cd your-project
 agentpack init --agent codex       # or claude, cursor, windsurf, antigravity
-agentpack pack --task "fix auth token expiry"
+printf '%s\n' "fix auth token expiry" > .agentpack/task.md
+agentpack pack
 ```
 
 This creates `.agentpack/` state, installs the requested agent integration, generates a ranked context pack, and writes the adapter output for that agent. For active local work, keep context fresh with:
@@ -437,7 +438,8 @@ Most users only need four commands:
 
 ```bash
 agentpack init --agent codex
-agentpack pack --task "describe the change"
+printf '%s\n' "describe the change" > .agentpack/task.md
+agentpack pack
 agentpack watch
 agentpack doctor --agent all
 ```
@@ -640,18 +642,21 @@ Summaries are built with parallel AST/regex analysis — no network, no tokens s
 
 ### `agentpack pack`
 
-Generate a context pack.
+Generate a context pack. Task text lives in `.agentpack/task.md`; inline task strings are no longer supported on `pack`. `--task auto` remains for old hooks and scripts, and is the default when the flag is omitted.
 
 ```bash
-agentpack pack --task "fix auth session bug"        # auto-detects your IDE
-agentpack pack --agent claude --task "fix auth bug" # explicit agent
-agentpack pack --workspace apps/web --task "fix web auth"
+printf '%s\n' "fix auth session bug" > .agentpack/task.md
+agentpack pack                                # auto-detects your IDE
+agentpack pack --agent claude                 # explicit agent
+agentpack pack --workspace apps/web
 
 # Only include changes since a git ref
-agentpack pack --task "review these changes" --since main
+printf '%s\n' "review these changes" > .agentpack/task.md
+agentpack pack --since main
 
 # Watch mode — re-packs on every file change
-agentpack pack --task "refactor auth" --session
+printf '%s\n' "refactor auth" > .agentpack/task.md
+agentpack pack --session
 ```
 
 Options:
@@ -659,7 +664,7 @@ Options:
 | Flag | Default | Description |
 |------|---------|-------------|
 | `--agent` | `auto` | Target agent (`auto` \| `claude` \| `cursor` \| `windsurf` \| `codex` \| `antigravity` \| `generic`). `auto` detects the active IDE from env and project files. |
-| `--task` | `auto` | Task description, or `auto` to infer from git |
+| `--task` | `auto` | Backward-compatible task source. Only `auto` is supported; write task text to `.agentpack/task.md`. |
 | `--mode` | `balanced` | Budget mode: `minimal`, `balanced`, `deep` |
 | `--budget` | 0 (uses config default 25000) | Token budget |
 | `--workspace` | — | Restrict packing to a monorepo workspace and write `.agentpack/workspaces/<workspace>/context.md` |
@@ -1389,7 +1394,7 @@ src/agentpack/
 
 ## Roadmap
 
-Next release target: **0.2.0 = benchmark + recall release**.
+Next release target: **0.3.0 = public benchmark expansion + npm publish hardening**.
 
 - Expand public source-checkout fixtures and publish reproducible `benchmark --sample-fixtures --compare --misses` output.
 - Raise recall on real historical tasks while keeping token precision healthy; target 60%+ recall, 50%+ token precision, and balanced packs under 25k tokens.

{agentpack_cli-0.2.0 → agentpack_cli-0.2.1}/README.md

@@ -5,7 +5,7 @@
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
 [![CI](https://github.com/vishal2612200/agentpack/actions/workflows/ci.yml/badge.svg)](https://github.com/vishal2612200/agentpack/actions/workflows/ci.yml)
 
-> **Status: alpha (v0.2.0).** Works, tested, used in real sessions. Python and JavaScript/TypeScript are the best-supported languages. Not yet validated across a wide range of repos. API may change before 1.0.
+> **Status: alpha (v0.2.1).** Works, tested, used in real sessions. Python and JavaScript/TypeScript are the best-supported languages. Not yet validated across a wide range of repos. API may change before 1.0.
 >
 > **Platform note:** macOS and Linux are fully supported. Windows support is not yet implemented (git hooks use POSIX shell; the Claude Code session hooks use `python3`/`rm -f`). Contributions welcome.
 
@@ -61,7 +61,8 @@ The npm package is a Node launcher around the Python implementation. It installs
 ```bash
 cd your-project
 agentpack init --agent codex       # or claude, cursor, windsurf, antigravity
-agentpack pack --task "fix auth token expiry"
+printf '%s\n' "fix auth token expiry" > .agentpack/task.md
+agentpack pack
 ```
 
 This creates `.agentpack/` state, installs the requested agent integration, generates a ranked context pack, and writes the adapter output for that agent. For active local work, keep context fresh with:
@@ -398,7 +399,8 @@ Most users only need four commands:
 
 ```bash
 agentpack init --agent codex
-agentpack pack --task "describe the change"
+printf '%s\n' "describe the change" > .agentpack/task.md
+agentpack pack
 agentpack watch
 agentpack doctor --agent all
 ```
@@ -601,18 +603,21 @@ Summaries are built with parallel AST/regex analysis — no network, no tokens s
 
 ### `agentpack pack`
 
-Generate a context pack.
+Generate a context pack. Task text lives in `.agentpack/task.md`; inline task strings are no longer supported on `pack`. `--task auto` remains for old hooks and scripts, and is the default when the flag is omitted.
 
 ```bash
-agentpack pack --task "fix auth session bug"        # auto-detects your IDE
-agentpack pack --agent claude --task "fix auth bug" # explicit agent
-agentpack pack --workspace apps/web --task "fix web auth"
+printf '%s\n' "fix auth session bug" > .agentpack/task.md
+agentpack pack                                # auto-detects your IDE
+agentpack pack --agent claude                 # explicit agent
+agentpack pack --workspace apps/web
 
 # Only include changes since a git ref
-agentpack pack --task "review these changes" --since main
+printf '%s\n' "review these changes" > .agentpack/task.md
+agentpack pack --since main
 
 # Watch mode — re-packs on every file change
-agentpack pack --task "refactor auth" --session
+printf '%s\n' "refactor auth" > .agentpack/task.md
+agentpack pack --session
 ```
 
 Options:
@@ -620,7 +625,7 @@ Options:
 | Flag | Default | Description |
 |------|---------|-------------|
 | `--agent` | `auto` | Target agent (`auto` \| `claude` \| `cursor` \| `windsurf` \| `codex` \| `antigravity` \| `generic`). `auto` detects the active IDE from env and project files. |
-| `--task` | `auto` | Task description, or `auto` to infer from git |
+| `--task` | `auto` | Backward-compatible task source. Only `auto` is supported; write task text to `.agentpack/task.md`. |
 | `--mode` | `balanced` | Budget mode: `minimal`, `balanced`, `deep` |
 | `--budget` | 0 (uses config default 25000) | Token budget |
 | `--workspace` | — | Restrict packing to a monorepo workspace and write `.agentpack/workspaces/<workspace>/context.md` |
@@ -1350,7 +1355,7 @@ src/agentpack/
 
 ## Roadmap
 
-Next release target: **0.2.0 = benchmark + recall release**.
+Next release target: **0.3.0 = public benchmark expansion + npm publish hardening**.
 
 - Expand public source-checkout fixtures and publish reproducible `benchmark --sample-fixtures --compare --misses` output.
 - Raise recall on real historical tasks while keeping token precision healthy; target 60%+ recall, 50%+ token precision, and balanced packs under 25k tokens.

{agentpack_cli-0.2.0 → agentpack_cli-0.2.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "agentpack-cli"
-version = "0.2.0"
+version = "0.2.1"
 description = "Task-aware context packing for AI coding agents — Claude, Cursor, Windsurf, Codex, and Antigravity"
 readme = "README.md"
 requires-python = ">=3.10"

{agentpack_cli-0.2.0 → agentpack_cli-0.2.1}/src/agentpack/__init__.py

@@ -1,3 +1,3 @@
 """AgentPack — task-aware context packing for AI coding agents."""
 
-__version__ = "0.2.0"
+__version__ = "0.2.1"

{agentpack_cli-0.2.0 → agentpack_cli-0.2.1}/src/agentpack/commands/doctor.py

@@ -127,7 +127,7 @@ def register(app: typer.Typer) -> None:
                 age_str = f"{int(age // 3600)}h {int((age % 3600) // 60)}m" if age > 3600 else f"{int(age // 60)}m"
                 console.print(f"  [green]✓[/] context pack present (age: {age_str})")
             else:
-                console.print("  [yellow]![/] No context pack yet — run: agentpack pack --task \"<task>\"")
+                console.print("  [yellow]![/] No context pack yet — write .agentpack/task.md, then run: agentpack pack --task auto")
 
         # --- Agent-specific config ---
         console.print("\n[bold]Agent config[/]")

{agentpack_cli-0.2.0 → agentpack_cli-0.2.1}/src/agentpack/commands/hook_cmd.py

@@ -306,8 +306,13 @@ def _run_user_prompt_submit(root: Path) -> None:
     should_repack = repo_changed or task_switched or pack_task_changed
 
     if should_repack:
+        if task != "auto":
+            try:
+                _write_task_md(root, task)
+            except Exception:
+                pass
         subprocess.Popen(
-            ["agentpack", "pack", "--task", task, "--mode", "balanced", "--since", "HEAD~1"],
+            ["agentpack", "pack", "--task", "auto", "--mode", "balanced", "--since", "HEAD~1"],
             stdout=subprocess.DEVNULL,
             stderr=subprocess.DEVNULL,
         )
@@ -361,7 +366,7 @@ def _run_user_prompt_submit(root: Path) -> None:
             )
         else:
             msg = (
-                "AgentPack active. Run `agentpack pack --task \"<task>\"` to build context.\n"
+                "AgentPack active. Write `.agentpack/task.md`, then run `agentpack pack --task auto` to build context.\n"
                 "For auto context, install MCP: agentpack install --agent claude"
             )
 

{agentpack_cli-0.2.0 → agentpack_cli-0.2.1}/src/agentpack/commands/install.py

@@ -37,7 +37,7 @@ def _validate_install_agent(agent: str) -> None:
 def _print_install_results(agent: str, results: dict[str, str]) -> None:
     if not results:
         console.print("[green]Generic agent selected.[/] No agent-specific hooks are required.")
-        console.print('  Run [bold]agentpack pack --agent generic --task "<task>"[/] to generate context.')
+        console.print("  Write [bold].agentpack/task.md[/], then run [bold]agentpack pack --agent generic --task auto[/] to generate context.")
         return
 
     for key, action in results.items():
@@ -53,7 +53,7 @@ def _print_install_results(agent: str, results: dict[str, str]) -> None:
             console.print(f"[green]{key} {action}.[/]")
 
     if agent in {"cursor", "windsurf", "codex", "generic"}:
-        console.print(f'  Run [bold]agentpack pack --agent {agent} --task "<task>"[/] to generate context.')
+        console.print(f"  Write [bold].agentpack/task.md[/], then run [bold]agentpack pack --agent {agent} --task auto[/] to generate context.")
     elif agent == "antigravity":
         console.print("  AgentPack Skill will activate automatically in Antigravity for coding tasks.")
 

{agentpack_cli-0.2.0 → agentpack_cli-0.2.1}/src/agentpack/commands/pack.py

@@ -21,7 +21,11 @@ def register(app: typer.Typer) -> None:
     @app.command()
     def pack(
         agent: str = typer.Option("auto", "--agent", help="Target agent (auto|claude|cursor|windsurf|codex|antigravity|generic). 'auto' detects from environment."),
-        task: str = typer.Option("auto", "--task", help="Task description, or 'auto' to infer from git."),
+        task: str = typer.Option(
+            "auto",
+            "--task",
+            help="Task source. Only 'auto' is supported; write the task to .agentpack/task.md.",
+        ),
         mode: str = typer.Option("balanced", "--mode", help="Budget mode (minimal|balanced|deep)."),
         budget: int = typer.Option(0, "--budget", help="Token budget (0 = use config default)."),
         workspace: str = typer.Option("", "--workspace", help="Restrict pack to a monorepo workspace, e.g. apps/web."),
@@ -74,7 +78,12 @@ def _resolve_task(task: str) -> str:
 
 def _resolve_task_with_source(task: str) -> tuple[str, str]:
     if task != "auto":
-        return task, "explicit"
+        console.print(
+            "[red]`agentpack pack --task \"...\"` is no longer supported.[/]\n"
+            "Write the task to [bold].agentpack/task.md[/], then run "
+            "[bold]agentpack pack --task auto[/] or [bold]agentpack pack[/]."
+        )
+        raise typer.Exit(2)
     root = _root()
     # task.md takes priority over all git heuristics
     task_md_path = root / ".agentpack" / "task.md"

{agentpack_cli-0.2.0 → agentpack_cli-0.2.1}/src/agentpack/data/agentpack.md

@@ -9,9 +9,8 @@ Pack repo context and immediately start working on the task.
 ## Usage
 
 ```
-/agentpack --task "fix Redis SSE cancellation issue"
-/agentpack --task "add rate limiting to auth endpoints" --mode deep
-/agentpack --task auto      # infer task from branch + changed files + git log
+/agentpack
+/agentpack --mode deep
 /agentpack init             # interactive: prompts for default mode
 /agentpack status
 /agentpack stats
@@ -43,7 +42,6 @@ To start a session:
 ```bash
 agentpack session start                     # creates session + generates initial context
 agentpack session start --agent claude      # specify agent
-agentpack session start --task "fix bug"    # set initial task
 
 agentpack watch                             # in another terminal — auto-refreshes on changes
 ```
@@ -73,7 +71,8 @@ Then use normal prompts — context stays current while `watch` is running.
 ## Manual Pack Mode (no session)
 
 ```bash
-agentpack pack --agent claude --task "<task>" --mode balanced
+printf '%s\n' "<task>" > .agentpack/task.md
+agentpack pack --agent claude --task auto --mode balanced
 ```
 
 Then read `.agentpack/context.claude.md` in full.
@@ -136,7 +135,6 @@ agentpack explain --omitted                            # see what was excluded a
 
 | User types | Action |
 |---|---|
-| `/agentpack --task "..."` | check session or pack + read + work |
 | `/agentpack` | check session or pack with `--task auto` + read + work |
 | `/agentpack session start` | `agentpack session start` |
 | `/agentpack session status` | `agentpack session status` |
@@ -155,7 +153,8 @@ agentpack explain --omitted                            # see what was excluded a
 ## Notes
 
 - All commands are local — no API calls
-- `--task auto` infers from branch name → changed file paths → recent commit
+- Task text belongs in `.agentpack/task.md`; inline task strings are not supported on `agentpack pack`.
+- `--task auto` reads `.agentpack/task.md`, then falls back to branch name → changed file paths → recent commit
 - Changed files are highest priority in context
 - Session context files: `.agentpack/context.md` (readable), `.agentpack/context.compact.md` (compact)
 - Never overwrite `.agentignore` or `config.toml` without `--force`

{agentpack_cli-0.2.0 → agentpack_cli-0.2.1}/src/agentpack/installers/claude.py

@@ -25,7 +25,8 @@ Other tools:
 If MCP is not available, fall back to the CLI:
 
 ```bash
-agentpack pack --agent claude --task "<task>"
+printf '%s\n' "<task>" > .agentpack/task.md
+agentpack pack --agent claude --task auto
 ```
 
 Then read `.agentpack/context.claude.md`.

{agentpack_cli-0.2.0 → agentpack_cli-0.2.1}/src/agentpack/installers/cursor.py

@@ -14,7 +14,7 @@ At the start of every coding task:
 3. Read `.agentpack/context.md`.
 4. Use files listed in context as starting points, but verify with actual code before editing.
 When the user switches to a different coding task, repeat steps 1-3 before editing.
-If context is missing: `agentpack pack --agent cursor --task "<task>"`
+If context is missing: write `.agentpack/task.md`, then run `agentpack pack --agent cursor --task auto`.
 <!-- agentpack:rule:end -->"""
 
 _RULE_RE = re.compile(
@@ -66,7 +66,7 @@ At the start of every coding task:
 
 When the user switches to a different coding task, repeat steps 1-3 before editing.
 
-If context is missing: `agentpack pack --agent cursor --task "<task>"`
+If context is missing: write `.agentpack/task.md`, then run `agentpack pack --agent cursor --task auto`.
 """
         already = mdc_path.exists()
         if already and mdc_path.read_text() == content:

{agentpack_cli-0.2.0 → agentpack_cli-0.2.1}/src/agentpack/installers/windsurf.py

@@ -14,7 +14,7 @@ At the start of every coding task:
 3. Read `.agentpack/context.md`.
 4. Use files listed in context as starting points, but verify with actual code before editing.
 When the user switches to a different coding task, repeat steps 1-3 before editing.
-If context is missing: `agentpack pack --agent windsurf --task "<task>"`
+If context is missing: write `.agentpack/task.md`, then run `agentpack pack --agent windsurf --task auto`.
 <!-- agentpack:rule:end -->"""
 
 _RULE_RE = re.compile(