spec-runner 2.1.0__tar.gz → 2.2.0__tar.gz

{spec_runner-2.1.0/src/spec_runner.egg-info → spec_runner-2.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: spec-runner
-Version: 2.1.0
+Version: 2.2.0
 Summary: Task automation from markdown specs via Claude CLI
 Author: Andrei
 License-Expression: MIT
@@ -369,6 +369,8 @@ paths:
 |-----|--------------|------------------|
 | Claude | Yes | `{cmd} -p {prompt} --model {model}` |
 | Codex | Yes | `{cmd} -p {prompt} --model {model}` |
+| OpenCode ([sst/opencode](https://opencode.ai)) | Yes | `{cmd} run --model {model} {prompt}` |
+| Pi Agent ([pi.dev](https://pi.dev)) | Yes (basename match) | `{cmd} -p --model {model} {prompt}` |
 | Ollama | Yes | `{cmd} run {model} {prompt}` |
 | llama-cli | Yes | `{cmd} -m {model} -p {prompt} --no-display-prompt` |
 | Custom | Use template | `{cmd} --prompt {prompt}` |

{spec_runner-2.1.0 → spec_runner-2.2.0}/README.md

@@ -334,6 +334,8 @@ paths:
 |-----|--------------|------------------|
 | Claude | Yes | `{cmd} -p {prompt} --model {model}` |
 | Codex | Yes | `{cmd} -p {prompt} --model {model}` |
+| OpenCode ([sst/opencode](https://opencode.ai)) | Yes | `{cmd} run --model {model} {prompt}` |
+| Pi Agent ([pi.dev](https://pi.dev)) | Yes (basename match) | `{cmd} -p --model {model} {prompt}` |
 | Ollama | Yes | `{cmd} run {model} {prompt}` |
 | llama-cli | Yes | `{cmd} -m {model} -p {prompt} --no-display-prompt` |
 | Custom | Use template | `{cmd} --prompt {prompt}` |

{spec_runner-2.1.0 → spec_runner-2.2.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "spec-runner"
-version = "2.1.0"
+version = "2.2.0"
 description = "Task automation from markdown specs via Claude CLI"
 readme = "README.md"
 requires-python = ">=3.10"

{spec_runner-2.1.0 → spec_runner-2.2.0}/src/spec_runner/audit.py

@@ -116,9 +116,7 @@ def audit_all(config: ExecutorConfig, *, strict: bool = False) -> AuditReport:
     """
     report = AuditReport(strict=strict)
 
-    tasks: list[Task] = (
-        parse_tasks(config.tasks_file) if config.tasks_file.exists() else []
-    )
+    tasks: list[Task] = parse_tasks(config.tasks_file) if config.tasks_file.exists() else []
 
     req_ids: set[str] = set()
     if config.requirements_file.exists():
@@ -168,8 +166,7 @@ def audit_all(config: ExecutorConfig, *, strict: bool = False) -> AuditReport:
                             category=CAT_DANGLING_DESIGN_REF,
                             subject=ref,
                             message=(
-                                f"{task.id} references {ref} but it is not "
-                                "defined in design.md"
+                                f"{task.id} references {ref} but it is not defined in design.md"
                             ),
                             location=task.id,
                         )
@@ -177,9 +174,7 @@ def audit_all(config: ExecutorConfig, *, strict: bool = False) -> AuditReport:
 
     # 4. Uncovered requirements — REQ defined but no task references it
     if req_ids:
-        covered_reqs = {
-            ref for task in tasks for ref in task.traces_to if ref.startswith("REQ-")
-        }
+        covered_reqs = {ref for task in tasks for ref in task.traces_to if ref.startswith("REQ-")}
         for req in sorted(req_ids - covered_reqs):
             report.findings.append(
                 AuditFinding(
@@ -194,10 +189,7 @@ def audit_all(config: ExecutorConfig, *, strict: bool = False) -> AuditReport:
     # 5. Uncovered designs — DESIGN defined but no task references it
     if design_ids:
         covered_designs = {
-            ref
-            for task in tasks
-            for ref in task.traces_to
-            if ref.startswith("DESIGN-")
+            ref for task in tasks for ref in task.traces_to if ref.startswith("DESIGN-")
         }
         for design in sorted(design_ids - covered_designs):
             report.findings.append(

{spec_runner-2.1.0 → spec_runner-2.2.0}/src/spec_runner/cli.py

@@ -5,6 +5,7 @@ import json
 import signal
 import sys
 import time
+from collections.abc import Callable
 from datetime import datetime
 from pathlib import Path
 from uuid import uuid4
@@ -622,7 +623,7 @@ def _dispatch_task_command(args: argparse.Namespace) -> None:
     tasks_file = Path(f"spec/{prefix}tasks.md") if prefix else TASKS_FILE
     tasks = parse_tasks(tasks_file)
 
-    write_commands = {
+    write_commands: dict[str, Callable[..., object]] = {
         "start": cmd_start,
         "done": cmd_done,
         "block": cmd_block,

{spec_runner-2.1.0 → spec_runner-2.2.0}/src/spec_runner/config.py

@@ -138,10 +138,13 @@ class ExecutorConfig:
     # Examples:
     #   claude: "{cmd} -p {prompt}" or "{cmd} -p {prompt} --model {model}"
     #   codex: "{cmd} -p {prompt}"
+    #   opencode: "{cmd} run --model {model} {prompt}"
+    #   pi: "{cmd} -p --model {model} {prompt}"
     #   ollama: "{cmd} run {model} {prompt}"
     #   llama-cli: "{cmd} -m {model} -p {prompt} --no-display-prompt"
     #   llama-server: "curl -s http://localhost:8080/completion -d '{{\"prompt\": {prompt}}}'"
-    # If empty, auto-detects based on command name
+    # If empty, auto-detects based on command name (claude, codex, opencode, pi,
+    # ollama, llama-cli, llama-server)
     command_template: str = ""
 
     # Hooks

{spec_runner-2.1.0 → spec_runner-2.2.0}/src/spec_runner/execution.py

@@ -375,7 +375,7 @@ def compute_retry_delay(error_code: ErrorCode | str, attempt: int, base_delay: i
     if strategy == "fatal":
         return 0.0
     if strategy == "backoff_exponential":
-        return min(30.0 * (2**attempt), 300.0)
+        return float(min(30.0 * (2**attempt), 300.0))
     return float(base_delay * (attempt + 1))
 
 
@@ -396,24 +396,13 @@ def _check_task_budget(
     """
     spent = state.task_cost(task_id)
     if config.task_budget_usd is not None and spent >= config.task_budget_usd:
-        return (
-            f"Task budget exceeded "
-            f"(${spent:.2f} >= ${config.task_budget_usd:.2f})"
-        )
+        return f"Task budget exceeded (${spent:.2f} >= ${config.task_budget_usd:.2f})"
 
-    if (
-        config.max_retry_cost_usd is not None
-        and attempt_index > 0
-    ):
+    if config.max_retry_cost_usd is not None and attempt_index > 0:
         ts = state.get_task_state(task_id)
-        retry_spent = (
-            sum(a.cost_usd or 0.0 for a in ts.attempts[1:]) if ts else 0.0
-        )
+        retry_spent = sum(a.cost_usd or 0.0 for a in ts.attempts[1:]) if ts else 0.0
         if retry_spent >= config.max_retry_cost_usd:
-            return (
-                f"Retry budget exceeded "
-                f"(${retry_spent:.2f} >= ${config.max_retry_cost_usd:.2f})"
-            )
+            return f"Retry budget exceeded (${retry_spent:.2f} >= ${config.max_retry_cost_usd:.2f})"
     return None
 
 

{spec_runner-2.1.0 → spec_runner-2.2.0}/src/spec_runner/github_sync.py

@@ -117,7 +117,7 @@ def _status_from_issue(issue: dict) -> str:
     if issue["state"] == "CLOSED":
         return "done"
     for label in issue.get("labels", []):
-        name = label["name"] if isinstance(label, dict) else label
+        name = str(label["name"]) if isinstance(label, dict) else str(label)
         if name.startswith("status:"):
             status = name.split(":", 1)[1]
             if status in STATUS_EMOJI:

{spec_runner-2.1.0 → spec_runner-2.2.0}/src/spec_runner/obs.py

@@ -17,7 +17,7 @@ from collections.abc import Iterator
 from contextlib import contextmanager
 from datetime import UTC, datetime
 from pathlib import Path
-from typing import Any
+from typing import Any, cast
 
 import structlog
 import ulid
@@ -190,7 +190,8 @@ def init_logging(
 
 
 def get_logger(module: str | None = None) -> structlog.BoundLogger:
-    return structlog.get_logger(module=module) if module else structlog.get_logger()
+    logger = structlog.get_logger(module=module) if module else structlog.get_logger()
+    return cast("structlog.BoundLogger", logger)
 
 
 class Span:

{spec_runner-2.1.0 → spec_runner-2.2.0}/src/spec_runner/report.py

@@ -50,11 +50,7 @@ class TraceabilityReport:
     @property
     def has_gaps(self) -> bool:
         """True when CI should flag this report as incomplete."""
-        return bool(
-            self.orphan_tasks
-            or self.uncovered_requirements
-            or self.unreferenced_designs
-        )
+        return bool(self.orphan_tasks or self.uncovered_requirements or self.unreferenced_designs)
 
 
 def _extract_section_ids(text: str, prefix: str) -> list[str]:
@@ -135,16 +131,12 @@ def build_report(
     # Gap warnings (LABS-42): identifiers defined in spec files that no
     # task references. Useful for CI integration — CI can fail if the
     # report has gaps, surfacing drift between specs and implementation.
-    report.uncovered_requirements = sorted(
-        req for req in all_reqs if req not in req_to_tasks
-    )
+    report.uncovered_requirements = sorted(req for req in all_reqs if req not in req_to_tasks)
     if design_to_req:
         referenced_designs = {
             ref for task in tasks for ref in task.traces_to if ref.startswith("DESIGN-")
         }
-        report.unreferenced_designs = sorted(
-            set(design_to_req) - referenced_designs
-        )
+        report.unreferenced_designs = sorted(set(design_to_req) - referenced_designs)
 
     # Build rows
     with ExecutorState(config) as state:

{spec_runner-2.1.0 → spec_runner-2.2.0}/src/spec_runner/runner.py

@@ -149,7 +149,8 @@ def build_cli_command(
     """Build CLI command from template or auto-detect based on command name.
 
     Args:
-        cmd: CLI command name (e.g., "claude", "codex", "llama-cli")
+        cmd: CLI command name (e.g., "claude", "codex", "opencode", "pi",
+            "ollama", "llama-cli")
         prompt: The prompt text
         model: Model name (optional)
         template: Command template with placeholders (optional)
@@ -182,6 +183,9 @@ def build_cli_command(
 
     # Auto-detect based on command name
     cmd_lower = cmd.lower()
+    # "pi" is too short for substring matching — match on basename only to
+    # avoid false positives like "/usr/local/bin/anti-pi" or "opencode-pi-cli".
+    cmd_basename = Path(cmd).name.lower()
 
     if "llama-cli" in cmd_lower or "llama.cpp" in cmd_lower:
         # llama.cpp CLI
@@ -199,6 +203,16 @@ def build_cli_command(
         # Ollama CLI
         return [cmd, "run", model or "llama3", prompt]
 
+    elif "opencode" in cmd_lower:
+        # sst/opencode: `opencode run [--model provider/id] <prompt>`
+        # Prompt is positional, model accepts "provider/model" form
+        # (e.g. "anthropic/claude-3-5-sonnet").
+        result = [cmd, "run"]
+        if model:
+            result.extend(["--model", model])
+        result.append(prompt)
+        return result
+
     elif "codex" in cmd_lower:
         # Codex CLI
         result = [cmd, "-p", prompt]
@@ -206,6 +220,17 @@ def build_cli_command(
             result.extend(["--model", model])
         return result
 
+    elif cmd_basename == "pi" or cmd_basename.startswith("pi."):
+        # earendil-works/pi: `pi -p [--model X] <prompt>` (non-interactive mode)
+        # Model accepts "provider/id" or bare model name; defaults driven by
+        # `~/.config/pi/config.yaml`. Match on basename to avoid short-name
+        # collisions (see cmd_basename comment above).
+        result = [cmd, "-p"]
+        if model:
+            result.extend(["--model", model])
+        result.append(prompt)
+        return result
+
     else:
         # Claude CLI (default)
         result = [cmd, "-p", prompt]
@@ -297,4 +322,4 @@ async def run_claude_async(
             proc.kill()
             await proc.wait()
         raise
-    return stdout_bytes.decode(), stderr_bytes.decode(), proc.returncode
+    return stdout_bytes.decode(), stderr_bytes.decode(), proc.returncode or 0

spec_runner-2.2.0/src/spec_runner/skills/spec-generator-skill/templates/prompts/review.opencode.md

@@ -0,0 +1,38 @@
+# Code Review
+
+Task: ${TASK_ID} — ${TASK_NAME}
+
+Files changed:
+${CHANGED_FILES}
+
+Diff:
+${GIT_DIFF}
+
+## Instructions
+
+Review the code for:
+1. Bugs and errors
+2. Security issues
+3. Missing error handling
+4. Test coverage
+
+## Required Response Format
+
+You MUST end your response with exactly one of these status codes on a new line:
+
+```
+REVIEW_PASSED
+```
+Use this if the code looks good and has no issues.
+
+```
+REVIEW_FIXED
+```
+Use this if you found and fixed issues.
+
+```
+REVIEW_FAILED
+```
+Use this if there are issues that need manual attention.
+
+Do not add any text after the status code.

spec_runner-2.2.0/src/spec_runner/skills/spec-generator-skill/templates/prompts/review.pi.md

@@ -0,0 +1,38 @@
+# Code Review
+
+Task: ${TASK_ID} — ${TASK_NAME}
+
+Files changed:
+${CHANGED_FILES}
+
+Diff:
+${GIT_DIFF}
+
+## Instructions
+
+Review the code for:
+1. Bugs and errors
+2. Security issues
+3. Missing error handling
+4. Test coverage
+
+## Required Response Format
+
+You MUST end your response with exactly one of these status codes on a new line:
+
+```
+REVIEW_PASSED
+```
+Use this if the code looks good and has no issues.
+
+```
+REVIEW_FIXED
+```
+Use this if you found and fixed issues.
+
+```
+REVIEW_FAILED
+```
+Use this if there are issues that need manual attention.
+
+Do not add any text after the status code.

{spec_runner-2.1.0 → spec_runner-2.2.0}/src/spec_runner/state.py

@@ -213,6 +213,7 @@ class ExecutorState:
         # Init DB first so tables exist
         self._init_db()
 
+        assert self._conn is not None
         with self._conn:
             # Migrate tasks and attempts
             for task_id, task_data in data.get("tasks", {}).items():
@@ -262,6 +263,7 @@ class ExecutorState:
 
     def _load(self) -> None:
         """Load state from SQLite into in-memory dicts."""
+        assert self._conn is not None
         # Load tasks
         cursor = self._conn.execute("SELECT task_id, status, started_at, completed_at FROM tasks")
         for row in cursor.fetchall():
@@ -323,6 +325,7 @@ class ExecutorState:
 
     def _save_meta(self) -> None:
         """Persist meta counters to SQLite."""
+        assert self._conn is not None
         for key, value in [
             ("consecutive_failures", str(self.consecutive_failures)),
             ("total_completed", str(self.total_completed)),
@@ -340,6 +343,7 @@ class ExecutorState:
         Called by external code (e.g. executor.py) when direct
         mutations are made to in-memory state outside record_attempt/mark_running.
         """
+        assert self._conn is not None
         with self._conn:
             # Upsert all tasks
             for task_id, ts in self.tasks.items():
@@ -415,6 +419,7 @@ class ExecutorState:
             review_findings=review_findings,
         )
         state.attempts.append(attempt)
+        assert self._conn is not None
 
         if success:
             state.status = "success"
@@ -519,6 +524,7 @@ class ExecutorState:
         state = self.get_task_state(task_id)
         state.status = "running"
         state.started_at = datetime.now().isoformat()
+        assert self._conn is not None
 
         try:
             with self._conn:

{spec_runner-2.1.0 → spec_runner-2.2.0}/src/spec_runner/tui.py

@@ -10,6 +10,7 @@ import contextlib
 import sqlite3
 from datetime import datetime
 from pathlib import Path
+from typing import Any
 
 from textual.app import App, ComposeResult
 from textual.binding import Binding
@@ -159,7 +160,7 @@ class StatsBar(Static):
 class KanbanColumn(Vertical):
     """A single column in the Kanban board."""
 
-    def __init__(self, title: str, **kwargs: object) -> None:
+    def __init__(self, title: str, **kwargs: Any) -> None:
         super().__init__(**kwargs)
         self.border_title = title
 
@@ -167,7 +168,7 @@ class KanbanColumn(Vertical):
 class LogPanel(Static):
     """Panel showing execution progress log, tailing a progress file."""
 
-    def __init__(self, **kwargs: object) -> None:
+    def __init__(self, **kwargs: Any) -> None:
         super().__init__(**kwargs)
         self._file_pos: int = 0
         self._lines: list[str] = []
@@ -553,6 +554,6 @@ class SpecRunnerApp(App[None]):
         else:
             log_panel.add_line(f"[bold cyan]While paused: {summary}[/bold cyan]")
 
-    def action_quit(self) -> None:
+    async def action_quit(self) -> None:
         """Quit the TUI."""
         self.exit()

{spec_runner-2.1.0 → spec_runner-2.2.0/src/spec_runner.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: spec-runner
-Version: 2.1.0
+Version: 2.2.0
 Summary: Task automation from markdown specs via Claude CLI
 Author: Andrei
 License-Expression: MIT
@@ -369,6 +369,8 @@ paths:
 |-----|--------------|------------------|
 | Claude | Yes | `{cmd} -p {prompt} --model {model}` |
 | Codex | Yes | `{cmd} -p {prompt} --model {model}` |
+| OpenCode ([sst/opencode](https://opencode.ai)) | Yes | `{cmd} run --model {model} {prompt}` |
+| Pi Agent ([pi.dev](https://pi.dev)) | Yes (basename match) | `{cmd} -p --model {model} {prompt}` |
 | Ollama | Yes | `{cmd} run {model} {prompt}` |
 | llama-cli | Yes | `{cmd} -m {model} -p {prompt} --no-display-prompt` |
 | Custom | Use template | `{cmd} --prompt {prompt}` |

{spec_runner-2.1.0 → spec_runner-2.2.0}/src/spec_runner.egg-info/SOURCES.txt

@@ -54,6 +54,8 @@ src/spec_runner/skills/spec-generator-skill/templates/prompts/review.codex.md
 src/spec_runner/skills/spec-generator-skill/templates/prompts/review.llama.md
 src/spec_runner/skills/spec-generator-skill/templates/prompts/review.md
 src/spec_runner/skills/spec-generator-skill/templates/prompts/review.ollama.md
+src/spec_runner/skills/spec-generator-skill/templates/prompts/review.opencode.md
+src/spec_runner/skills/spec-generator-skill/templates/prompts/review.pi.md
 tests/test_audit.py
 tests/test_audit_log.py
 tests/test_config.py

{spec_runner-2.1.0 → spec_runner-2.2.0}/tests/test_audit.py

@@ -177,9 +177,7 @@ class TestUncoveredSpec:
         config = _write_specs(tmp_path, tasks, REQS_CLEAN, DESIGN_CLEAN)
         report = audit_all(config)
 
-        uncovered_designs = [
-            f for f in report.findings if f.category == CAT_UNCOVERED_DESIGN
-        ]
+        uncovered_designs = [f for f in report.findings if f.category == CAT_UNCOVERED_DESIGN]
         assert {f.subject for f in uncovered_designs} == {"DESIGN-002"}
 
 

{spec_runner-2.1.0 → spec_runner-2.2.0}/tests/test_audit_log.py

@@ -155,9 +155,7 @@ class TestBuildAuditLogger:
         assert logger.operator == "maestro"
 
     def test_spec_prefix_flows_through(self, tmp_path: Path) -> None:
-        config = _make_config(
-            tmp_path, audit_log_path="audit.jsonl", spec_prefix="phase5-"
-        )
+        config = _make_config(tmp_path, audit_log_path="audit.jsonl", spec_prefix="phase5-")
         logger = build_audit_logger(config)
         logger.record(EVENT_RUN_STARTED)
         entry = _read_audit(logger.path)[0]
@@ -236,9 +234,7 @@ class TestExecutorStateAuditIntegration:
         assert failed["details"]["last_error"] == "boom"
         assert failed["details"]["error_code"] == "TASK_FAILED"
 
-    def test_degraded_mode_emits_state_degraded_event(
-        self, tmp_path: Path, monkeypatch
-    ) -> None:
+    def test_degraded_mode_emits_state_degraded_event(self, tmp_path: Path, monkeypatch) -> None:
         import sqlite3
         from unittest.mock import MagicMock
 

{spec_runner-2.1.0 → spec_runner-2.2.0}/tests/test_runner.py

@@ -50,6 +50,42 @@ class TestBuildCliCommand:
         assert "--model" in result
         assert "gpt-4" in result
 
+    def test_opencode_auto_detect(self):
+        result = build_cli_command("opencode", "hello")
+        assert result == ["opencode", "run", "hello"]
+
+    def test_opencode_with_model(self):
+        result = build_cli_command("opencode", "hello", model="anthropic/claude-sonnet-4-6")
+        assert result == [
+            "opencode",
+            "run",
+            "--model",
+            "anthropic/claude-sonnet-4-6",
+            "hello",
+        ]
+
+    def test_pi_auto_detect(self):
+        result = build_cli_command("pi", "hello")
+        assert result == ["pi", "-p", "hello"]
+
+    def test_pi_with_model(self):
+        result = build_cli_command("pi", "hello", model="openai/gpt-4o")
+        assert result == ["pi", "-p", "--model", "openai/gpt-4o", "hello"]
+
+    def test_pi_path_basename_match(self):
+        # Absolute path with pi as the basename should still auto-detect.
+        result = build_cli_command("/usr/local/bin/pi", "hello")
+        assert result == ["/usr/local/bin/pi", "-p", "hello"]
+
+    def test_pi_no_false_positive_substring(self):
+        # "pipe-cli" or anything containing "pi" should NOT be treated as Pi —
+        # it must fall through to the Claude default.
+        result = build_cli_command("pipe-cli", "hello")
+        assert result[0] == "pipe-cli"
+        # Claude default uses -p too, but key signal: prompt is the third
+        # arg ("-p hello"), not the fourth ("-p" + appended prompt).
+        assert result == ["pipe-cli", "-p", "hello"]
+
     def test_ollama_auto_detect(self):
         result = build_cli_command("ollama", "hello", model="llama3")
         assert result == ["ollama", "run", "llama3", "hello"]