simplicio-loop 3.38.0__tar.gz → 3.38.1__tar.gz

{simplicio_loop-3.38.0 → simplicio_loop-3.38.1}/LICENSE

@@ -1,21 +1,21 @@
-MIT License
-
-Copyright (c) 2026 Wesley Simplicio
-
-Permission is hereby granted, free of charge, to any person obtaining a copy
-of this software and associated documentation files (the "Software"), to deal
-in the Software without restriction, including without limitation the rights
-to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all
-copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
+MIT License
+
+Copyright (c) 2026 Wesley Simplicio
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

{simplicio_loop-3.38.0/simplicio_loop.egg-info → simplicio_loop-3.38.1}/PKG-INFO

@@ -1,83 +1,87 @@
-Metadata-Version: 2.4
-Name: simplicio-loop
-Version: 3.38.0
-Summary: The Universal Looping AI Orchestrator — a runtime-agnostic super-plugin (7 skills) that drains any queue of work end-to-end on any LLM/runtime.
-Author-email: Wesley Simplicio <wesleybob4@gmail.com>
-License: MIT
-Project-URL: Homepage, https://github.com/wesleysimplicio/simplicio-loop
-Project-URL: Repository, https://github.com/wesleysimplicio/simplicio-loop
-Project-URL: Issues, https://github.com/wesleysimplicio/simplicio-loop/issues
-Keywords: ai,agent,orchestrator,llm,claude,cursor,automation,ralph-loop,token-economy,cli
-Classifier: Development Status :: 4 - Beta
-Classifier: Intended Audience :: Developers
-Classifier: License :: OSI Approved :: MIT License
-Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3 :: Only
-Classifier: Topic :: Software Development :: Build Tools
-Classifier: Topic :: Utilities
-Classifier: Operating System :: OS Independent
-Requires-Python: >=3.8
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: simplicio-cli>=0.16.1
-Provides-Extra: dev
-Requires-Dist: pytest>=7; extra == "dev"
-Requires-Dist: pytest-cov>=4; extra == "dev"
-Provides-Extra: ml
-Requires-Dist: model2vec>=0.3; extra == "ml"
-Requires-Dist: numpy>=1.21; extra == "ml"
-Dynamic: license-file
-
-# simplicio-loop
-
-**The Universal Looping AI Orchestrator** — a runtime-agnostic **super-plugin** (7 skills) that
-drains any queue of work end-to-end on **any LLM / runtime**:
-`discover → implement → verify → merge → close → watch 24/7`, behind safety gates and evidence
-checks, at up to **96% fewer tokens**. Not a chatbot. A worker.
-
-![simplicio-loop](https://raw.githubusercontent.com/wesleysimplicio/simplicio-loop/main/assets/simplicio-loop-hero-2026.png)
-
-## Install
-
-```bash
-pip install simplicio-loop
-```
-
-Then drop the skills + hooks into your project (or globally):
-
-```bash
-simplicio-loop install            # into ./.claude of the current project
-simplicio-loop install --global   # into ~/.claude (all projects)
-```
-
-Now invoke it from your agent runtime (Claude Code, Cursor, Codex, Gemini, …):
-
-```
-/simplicio-loop finish all the open issues
-```
-
-## What you get — 7 skills
-
-| Skill | What it does |
-|---|---|
-| `simplicio-loop` | Unified public entrypoint: orchestrator core + hardened loop behind one command. |
-| `simplicio-tasks` | Legacy alias kept only for compatibility with older installs and saved prompts. |
-| `simplicio-orient` | Terminal-first token economy — output-reduction catalog, tee-cache, signatures-read. |
-| `simplicio-review` | Adversarial review — parallel subagents on distinct rubrics, deduped into one verdict. |
-| `simplicio-compress` | Output + memory compression, byte-preserving identifiers. |
-| `simplicio-autoresearch` | Evolutionary mutate/eval/keep-revert optimizer — yool-guardrailed caps, git-isolated branch, anti-Goodhart gate-first eval. |
-| `simplicio-learn` | Retrospective — durable, deduped lessons written back to memory. |
-
-## Highlights
-
-- **12 runtimes, one protocol** — Claude Code, Codex, VS Code/Copilot, Cursor, Antigravity, Kiro,
-  OpenCode, Gemini, Aider, Simplicio Agent (formerly Hermes), OpenClaw, Orca.
-- **Evidence-gated completion** — never a false "done"; exits only on a verified `<promise>`,
-  cap, spindle handoff, or STOP.
-- **Token economy** — honest "answer concisely" baseline; savings credited only on verified-correct
-  outcomes.
-
-Requires Python 3.8+. The skills, hooks, and installer are pure cross-platform Python.
-
-MIT — part of the [Simplicio](https://github.com/wesleysimplicio) ecosystem.
-Full docs: <https://github.com/wesleysimplicio/simplicio-loop>
+Metadata-Version: 2.4
+Name: simplicio-loop
+Version: 3.38.1
+Summary: The Universal Looping AI Orchestrator — a runtime-agnostic super-plugin (7 skills) that drains any queue of work end-to-end on any LLM/runtime.
+Author-email: Wesley Simplicio <wesleybob4@gmail.com>
+License: MIT
+Project-URL: Homepage, https://github.com/wesleysimplicio/simplicio-loop
+Project-URL: Repository, https://github.com/wesleysimplicio/simplicio-loop
+Project-URL: Issues, https://github.com/wesleysimplicio/simplicio-loop/issues
+Keywords: ai,agent,orchestrator,llm,claude,cursor,automation,ralph-loop,token-economy,cli
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Topic :: Software Development :: Build Tools
+Classifier: Topic :: Utilities
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: simplicio-cli>=0.16.2
+Requires-Dist: tiktoken<1,>=0.12.0
+Provides-Extra: dev
+Requires-Dist: pytest>=7; extra == "dev"
+Requires-Dist: pytest-cov>=4; extra == "dev"
+Requires-Dist: cryptography<50,>=42; extra == "dev"
+Provides-Extra: ml
+Requires-Dist: model2vec>=0.3; extra == "ml"
+Requires-Dist: numpy>=1.21; extra == "ml"
+Provides-Extra: uvloop
+Requires-Dist: uvloop>=0.19; platform_system != "Windows" and extra == "uvloop"
+Dynamic: license-file
+
+# simplicio-loop
+
+**The Universal Looping AI Orchestrator** — a runtime-agnostic **super-plugin** (7 skills) that
+drains any queue of work end-to-end on **any LLM / runtime**:
+`discover → implement → verify → merge → close → watch 24/7`, behind safety gates and evidence
+checks, at up to **96% fewer tokens**. Not a chatbot. A worker.
+
+![simplicio-loop](https://raw.githubusercontent.com/wesleysimplicio/simplicio-loop/main/assets/simplicio-loop-hero-2026.png)
+
+## Install
+
+```bash
+pip install simplicio-loop
+```
+
+Then drop the skills + hooks into your project (or globally):
+
+```bash
+simplicio-loop install            # into ./.claude of the current project
+simplicio-loop install --global   # into ~/.claude (all projects)
+```
+
+Now invoke it from your agent runtime (Claude Code, Cursor, Codex, Gemini, …):
+
+```
+/simplicio-loop finish all the open issues
+```
+
+## What you get — 7 skills
+
+| Skill | What it does |
+|---|---|
+| `simplicio-loop` | Unified public entrypoint: orchestrator core + hardened loop behind one command. |
+| `simplicio-tasks` | Legacy alias kept only for compatibility with older installs and saved prompts. |
+| `simplicio-orient` | Terminal-first token economy — output-reduction catalog, tee-cache, signatures-read. |
+| `simplicio-review` | Adversarial review — parallel subagents on distinct rubrics, deduped into one verdict. |
+| `simplicio-compress` | Output + memory compression, byte-preserving identifiers. |
+| `simplicio-autoresearch` | Evolutionary mutate/eval/keep-revert optimizer — yool-guardrailed caps, git-isolated branch, anti-Goodhart gate-first eval. |
+| `simplicio-learn` | Retrospective — durable, deduped lessons written back to memory. |
+
+## Highlights
+
+- **12 runtimes, one protocol** — Claude Code, Codex, VS Code/Copilot, Cursor, Antigravity, Kiro,
+  OpenCode, Gemini, Aider, Simplicio Agent (formerly Hermes), OpenClaw, Orca.
+- **Evidence-gated completion** — never a false "done"; exits only on a verified `<promise>`,
+  cap, spindle handoff, or STOP.
+- **Token economy** — honest "answer concisely" baseline; savings credited only on verified-correct
+  outcomes.
+
+Requires Python 3.8+. The skills, hooks, and installer are pure cross-platform Python.
+
+MIT — part of the [Simplicio](https://github.com/wesleysimplicio) ecosystem.
+Full docs: <https://github.com/wesleysimplicio/simplicio-loop>

{simplicio_loop-3.38.0 → simplicio_loop-3.38.1}/PYPI.md

@@ -1,53 +1,53 @@
-# simplicio-loop
-
-**The Universal Looping AI Orchestrator** — a runtime-agnostic **super-plugin** (7 skills) that
-drains any queue of work end-to-end on **any LLM / runtime**:
-`discover → implement → verify → merge → close → watch 24/7`, behind safety gates and evidence
-checks, at up to **96% fewer tokens**. Not a chatbot. A worker.
-
-![simplicio-loop](https://raw.githubusercontent.com/wesleysimplicio/simplicio-loop/main/assets/simplicio-loop-hero-2026.png)
-
-## Install
-
-```bash
-pip install simplicio-loop
-```
-
-Then drop the skills + hooks into your project (or globally):
-
-```bash
-simplicio-loop install            # into ./.claude of the current project
-simplicio-loop install --global   # into ~/.claude (all projects)
-```
-
-Now invoke it from your agent runtime (Claude Code, Cursor, Codex, Gemini, …):
-
-```
-/simplicio-loop finish all the open issues
-```
-
-## What you get — 7 skills
-
-| Skill | What it does |
-|---|---|
-| `simplicio-loop` | Unified public entrypoint: orchestrator core + hardened loop behind one command. |
-| `simplicio-tasks` | Legacy alias kept only for compatibility with older installs and saved prompts. |
-| `simplicio-orient` | Terminal-first token economy — output-reduction catalog, tee-cache, signatures-read. |
-| `simplicio-review` | Adversarial review — parallel subagents on distinct rubrics, deduped into one verdict. |
-| `simplicio-compress` | Output + memory compression, byte-preserving identifiers. |
-| `simplicio-autoresearch` | Evolutionary mutate/eval/keep-revert optimizer — yool-guardrailed caps, git-isolated branch, anti-Goodhart gate-first eval. |
-| `simplicio-learn` | Retrospective — durable, deduped lessons written back to memory. |
-
-## Highlights
-
-- **12 runtimes, one protocol** — Claude Code, Codex, VS Code/Copilot, Cursor, Antigravity, Kiro,
-  OpenCode, Gemini, Aider, Simplicio Agent (formerly Hermes), OpenClaw, Orca.
-- **Evidence-gated completion** — never a false "done"; exits only on a verified `<promise>`,
-  cap, spindle handoff, or STOP.
-- **Token economy** — honest "answer concisely" baseline; savings credited only on verified-correct
-  outcomes.
-
-Requires Python 3.8+. The skills, hooks, and installer are pure cross-platform Python.
-
-MIT — part of the [Simplicio](https://github.com/wesleysimplicio) ecosystem.
-Full docs: <https://github.com/wesleysimplicio/simplicio-loop>
+# simplicio-loop
+
+**The Universal Looping AI Orchestrator** — a runtime-agnostic **super-plugin** (7 skills) that
+drains any queue of work end-to-end on **any LLM / runtime**:
+`discover → implement → verify → merge → close → watch 24/7`, behind safety gates and evidence
+checks, at up to **96% fewer tokens**. Not a chatbot. A worker.
+
+![simplicio-loop](https://raw.githubusercontent.com/wesleysimplicio/simplicio-loop/main/assets/simplicio-loop-hero-2026.png)
+
+## Install
+
+```bash
+pip install simplicio-loop
+```
+
+Then drop the skills + hooks into your project (or globally):
+
+```bash
+simplicio-loop install            # into ./.claude of the current project
+simplicio-loop install --global   # into ~/.claude (all projects)
+```
+
+Now invoke it from your agent runtime (Claude Code, Cursor, Codex, Gemini, …):
+
+```
+/simplicio-loop finish all the open issues
+```
+
+## What you get — 7 skills
+
+| Skill | What it does |
+|---|---|
+| `simplicio-loop` | Unified public entrypoint: orchestrator core + hardened loop behind one command. |
+| `simplicio-tasks` | Legacy alias kept only for compatibility with older installs and saved prompts. |
+| `simplicio-orient` | Terminal-first token economy — output-reduction catalog, tee-cache, signatures-read. |
+| `simplicio-review` | Adversarial review — parallel subagents on distinct rubrics, deduped into one verdict. |
+| `simplicio-compress` | Output + memory compression, byte-preserving identifiers. |
+| `simplicio-autoresearch` | Evolutionary mutate/eval/keep-revert optimizer — yool-guardrailed caps, git-isolated branch, anti-Goodhart gate-first eval. |
+| `simplicio-learn` | Retrospective — durable, deduped lessons written back to memory. |
+
+## Highlights
+
+- **12 runtimes, one protocol** — Claude Code, Codex, VS Code/Copilot, Cursor, Antigravity, Kiro,
+  OpenCode, Gemini, Aider, Simplicio Agent (formerly Hermes), OpenClaw, Orca.
+- **Evidence-gated completion** — never a false "done"; exits only on a verified `<promise>`,
+  cap, spindle handoff, or STOP.
+- **Token economy** — honest "answer concisely" baseline; savings credited only on verified-correct
+  outcomes.
+
+Requires Python 3.8+. The skills, hooks, and installer are pure cross-platform Python.
+
+MIT — part of the [Simplicio](https://github.com/wesleysimplicio) ecosystem.
+Full docs: <https://github.com/wesleysimplicio/simplicio-loop>

{simplicio_loop-3.38.0 → simplicio_loop-3.38.1}/README.md

@@ -10,7 +10,7 @@
   <a href="#-the-7-skills--5-accelerators"><img src="https://img.shields.io/badge/skills-7-7C3AED" alt="7 skills"></a>
   <a href="#-source-adapters"><img src="https://img.shields.io/badge/source%20adapters-5-00E08A" alt="5 source adapters"></a>
   <a href="#-15-runtimes-one-protocol"><img src="https://img.shields.io/badge/runtimes-15%20(3%20garantidos%2B12%20best--effort)-2563EB" alt="15 runtimes (3 guaranteed + 12 best-effort)"></a>
-  <a href="#-the-49-extension-points"><img src="https://img.shields.io/badge/extension%20points-49-00E08A" alt="49 extension points"></a>
+  <a href="#-the-49-extension-points"><img src="https://img.shields.io/badge/extension%20points-50-00E08A" alt="50 extension points"></a>
   <a href="#-token-economy"><img src="https://img.shields.io/badge/savings-unverified-888888" alt="Savings — unverified"></a>
   <a href="LICENSE"><img src="https://img.shields.io/badge/license-MIT-blue" alt="License"></a>
   <a href="https://discord.gg/wM6tr7xVb"><img src="https://img.shields.io/badge/Discord-Join%20Simplicio-5865F2?logo=discord&logoColor=white" alt="Join the Simplicio Discord"></a>

{simplicio_loop-3.38.0 → simplicio_loop-3.38.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "simplicio-loop"
-version = "3.38.0"
+version = "3.38.1"
 description = "The Universal Looping AI Orchestrator — a runtime-agnostic super-plugin (7 skills) that drains any queue of work end-to-end on any LLM/runtime."
 readme = "PYPI.md"
 requires-python = ">=3.8"
@@ -24,17 +24,18 @@ classifiers = [
 # The simplicio-loop drive requires the operator package (`simplicio-cli`), which exposes
 # `simplicio-dev-cli` and also brings the survey binary `simplicio-mapper` transitively.
 dependencies = [
-  "simplicio-cli>=0.16.1",
+  "simplicio-cli>=0.16.2",
+  "tiktoken>=0.12.0,<1",
 ]
 
 # Optional: the real embedding backend for `simplicio-cli semantic --ml` / `simplicio-cli rag --ml`.
 # The native engine itself is stdlib-only; this only enables embedding-based semantic dedup +
 # meaning retrieval. `pip install "simplicio-loop[ml]"`.
 [project.optional-dependencies]
-# Dev/test: the suite ALSO self-runs on bare python3 (`python3 scripts/check.py`), so pytest is
-# only a convenience for nicer output / IDE integration — never required.
-dev = ["pytest>=7", "pytest-cov>=4"]
+# Dev/test dependencies used by the local quality gate and secure-transport fixtures.
+dev = ["pytest>=7", "pytest-cov>=4", "cryptography>=42,<50"]
 ml = ["model2vec>=0.3", "numpy>=1.21"]
+uvloop = ["uvloop>=0.19; platform_system != 'Windows'"]
 
 [project.urls]
 Homepage = "https://github.com/wesleysimplicio/simplicio-loop"
@@ -43,11 +44,15 @@ Issues = "https://github.com/wesleysimplicio/simplicio-loop/issues"
 
 [project.scripts]
 simplicio-loop = "simplicio_loop.cli:main"
+simplicio-hub = "simplicio_loop.hub_daemon:main"
 # issue #286 step 11: the multi-device remote-worker protocol's server + worker + supervisor
 # are genuinely installable binaries, not source that only runs from a git checkout.
 simplicio-remote-queue-server = "simplicio_loop.remote_queue_server_cli:main"
 simplicio-remote-worker = "simplicio_loop.remote_worker_cli:main"
 simplicio-remote-worker-supervisor = "simplicio_loop.remote_worker_supervisor_cli:main"
+# issue #516: observability/enforcement CLI for the #498 process-supervisor contract
+# (status/top/queue/cancel/drain/reports over the #514 ProcessSpec/Lease/Result adapter).
+simplicio-process-supervisor = "simplicio_loop.process_enforcement_cli:main"
 
 [tool.setuptools]
 packages = ["simplicio_loop"]
@@ -67,3 +72,9 @@ exclude_lines = [
   "pragma: no cover",
   "if __name__ == .__main__.:",
 ]
+
+[tool.pytest.ini_options]
+addopts = ["--import-mode=importlib"]
+# The bundle/plugin trees mirror source files and are validated by mirror-parity.
+# Collect the canonical tests once instead of executing mirrored probes repeatedly.
+testpaths = ["tests"]

{simplicio_loop-3.38.0 → simplicio_loop-3.38.1}/setup.cfg

@@ -1,4 +1,4 @@
-[egg_info]
-tag_build = 
-tag_date = 0
-
+[egg_info]
+tag_build = 
+tag_date = 0
+

{simplicio_loop-3.38.0 → simplicio_loop-3.38.1}/simplicio_loop/__init__.py

@@ -13,6 +13,6 @@ try:
     try:
         __version__ = _v("simplicio-loop")
     except PackageNotFoundError:
-        __version__ = "3.38.0"
+        __version__ = "3.38.1"
 except Exception:  # pragma: no cover
-        __version__ = "3.38.0"
+        __version__ = "3.38.1"

{simplicio_loop-3.38.0 → simplicio_loop-3.38.1}/simplicio_loop/_bundle/hooks/action_gate.py

@@ -149,25 +149,36 @@ def scan_secret_text(text):
     return hits
 
 
-def _staged_diff():
-    # Scan the CURRENT working repo (where the command runs), NOT where this script lives —
-    # installed as a hook in another project, the user's repo is the cwd.
-    r = _run(["git", "diff", "--cached", "--unified=0"], cwd=os.getcwd())
+def _effective_command_cwd(cmd):
+    """Resolve an explicit leading cd or git -C target for PreToolUse scans."""
+    base = os.getcwd()
+    candidate = None
+    leading = re.match(
+        r"""^\s*(?:cd|Set-Location)\s+(?:"([^"]+)"|'([^']+)'|([^&;\n]+))\s*(?:&&|;)""",
+        cmd or "",
+    )
+    if leading:
+        candidate = next((value for value in leading.groups() if value), "").strip()
+    else:
+        git_c = re.search(r"""\bgit\s+-C\s+(?:"([^"]+)"|'([^']+)'|(\S+))""", cmd or "")
+        if git_c:
+            candidate = next((value for value in git_c.groups() if value), "").strip()
+    if not candidate:
+        return os.path.abspath(base)
+    candidate = os.path.expandvars(os.path.expanduser(candidate))
+    if not os.path.isabs(candidate):
+        candidate = os.path.join(base, candidate)
+    return os.path.normpath(os.path.abspath(candidate))
+
+
+def _staged_diff(cwd=None):
+    r = _run(["git", "diff", "--cached", "--unified=0"], cwd=cwd or os.getcwd())
     return (r.stdout if r and r.returncode == 0 else None)
 
 
-def _push_diff():
-    """Diff of the commits actually about to leave the machine on `git push` (#291).
-
-    The previously documented pre-push wiring (`action_gate.py check --staged`) scans the
-    STAGED diff — at push time that's almost always empty, since everything being pushed is
-    already committed. This scans the real push range instead: HEAD vs. the upstream tracking
-    branch when one is configured, falling back to the last commit (or the empty tree, for a
-    brand-new single-commit repo) when there is no upstream yet. Returns None only when no
-    range can be resolved at all — the caller treats that as fail-closed, same as an unreadable
-    staged diff.
-    """
-    cwd = os.getcwd()
+def _push_diff(cwd=None):
+    """Return the commits actually leaving the selected repository on git push."""
+    cwd = cwd or os.getcwd()
     r = _run(["git", "rev-parse", "--abbrev-ref", "--symbolic-full-name", "@{u}"], cwd=cwd)
     upstream = r.stdout.strip() if r and r.returncode == 0 and r.stdout.strip() else None
     if upstream:
@@ -177,16 +188,93 @@ def _push_diff():
     d = _run(["git", "diff", "HEAD~1..HEAD", "--unified=0"], cwd=cwd)
     if d and d.returncode == 0:
         return d.stdout
-    # No HEAD~1 (single-commit repo) — diff against git's well-known empty-tree hash.
     d = _run(["git", "diff", "4b825dc642cb6eb9a060e54bf8d69288fbee4904", "HEAD", "--unified=0"],
              cwd=cwd)
     return (d.stdout if d and d.returncode == 0 else None)
 
 
+def _delivery_contract(cwd):
+    """Load the optional frozen delivery contract for the effective repository."""
+    local_anchor = os.path.join(cwd, ".orchestrator", "loop", "anchor.json")
+    anchor_path = local_anchor if os.path.exists(local_anchor) else (
+        os.environ.get("SIMPLICIO_ANCHOR_FILE") or local_anchor
+    )
+    try:
+        with open(anchor_path, encoding="utf-8") as handle:
+            anchor = json.load(handle)
+        contract = anchor.get("delivery")
+        if not contract:
+            return None, None
+        if REPO not in sys.path:
+            sys.path.insert(0, REPO)
+        from simplicio_loop.delivery_contract import normalize_contract
+        return normalize_contract(contract), None
+    except FileNotFoundError:
+        return None, None
+    except (OSError, ValueError, TypeError, json.JSONDecodeError) as exc:
+        return None, "invalid frozen delivery contract: %s" % exc
+
+
+def _diff_new_paths(cwd, push):
+    """Return added paths from the same range that will be secret-scanned."""
+    if push:
+        upstream = _run(["git", "rev-parse", "--abbrev-ref", "--symbolic-full-name", "@{u}"], cwd=cwd)
+        ref = upstream.stdout.strip() if upstream and upstream.returncode == 0 else ""
+        ranges = (["%s...HEAD" % ref] if ref else ["HEAD~1..HEAD", "4b825dc642cb6eb9a060e54bf8d69288fbee4904", "HEAD"])
+        for item in ranges:
+            args = ["git", "diff", item, "--name-status"] if "..." in item or ".." in item else ["git", "diff", item, "HEAD", "--name-status"]
+            result = _run(args, cwd=cwd)
+            if result and result.returncode == 0:
+                return [line.split("\t", 1)[-1] for line in result.stdout.splitlines()
+                        if line and line[0] in {"A", "C"} and "\t" in line]
+        return []
+    result = _run(["git", "diff", "--cached", "--name-status"], cwd=cwd)
+    if not result or result.returncode != 0:
+        return []
+    return [line.split("\t", 1)[-1] for line in result.stdout.splitlines()
+            if line and line[0] in {"A", "C"} and "\t" in line]
+
+
+def _delivery_guard(cwd, diff, push):
+    contract, error = _delivery_contract(cwd)
+    if error:
+        return error
+    if contract is None:
+        return None
+    new_paths = _diff_new_paths(cwd, push)
+    if not contract["allow_new_files_in_repo"] and new_paths:
+        return "delivery contract forbids new files (cwd=%s): %s" % (cwd, ", ".join(sorted(new_paths)))
+    if not contract["allow_comments_in_code"]:
+        code_paths = {
+            line[6:].strip()
+            for line in diff.splitlines()
+            if line.startswith("+++ b/")
+        }
+        code_suffixes = (".cs", ".js", ".jsx", ".ts", ".tsx", ".py")
+        if not any(path.lower().endswith(code_suffixes) for path in code_paths):
+            return None
+        comments = []
+        for line in diff.splitlines():
+            if not line.startswith("+") or line.startswith("+++"):
+                continue
+            added = line[1:].lstrip()
+            if added.startswith(("#", "//", "/*", "*", '"""')):
+                comments.append(added[:80])
+        if comments:
+            return "delivery contract forbids new code comments (cwd=%s; lines=%d)" % (cwd, len(comments))
+    return None
+
+
 def _is_commit_or_push(cmd):
     return bool(re.search(r"\bgit\s+(commit|push)\b", cmd or ""))
 
 
+def _is_push(cmd):
+    return bool(re.search(r"\bgit\s+.*\bpush\b", cmd or ""))
+
+
+
+
 def _verdict(allow, reason=""):
     return {"action": "allow" if allow else "block", "reason": reason}
 
@@ -243,24 +331,35 @@ def _runtime_gate_escalation(cmd):
 
 
 def gate_command(cmd, staged=False):
-    """The core decision. Returns a verdict dict; BLOCK is fail-closed."""
-    # 1) irreversible op → block
+    """The core decision. Secret scans use the command's effective repository."""
     reason = classify_command(cmd)
     if reason:
         return _verdict(False, "irreversible op: " + reason)
-    # 2) a commit/push → secret-scan the staged diff (fail-closed if it can't be read)
     if staged or _is_commit_or_push(cmd):
-        diff = _staged_diff()
+        cwd = _effective_command_cwd(cmd)
+        push = _is_push(cmd)
+        diff = _push_diff(cwd) if push else _staged_diff(cwd)
         if diff is None:
-            # security check could not run on a push/commit → do not pass it
-            if _is_commit_or_push(cmd) or staged:
-                return _verdict(False, "cannot read staged diff to secret-scan — blocking the "
-                                       "commit/push (fail-closed). Stage changes or run in a git repo.")
-            return _verdict(True)
+            target = "push" if push else "staged"
+            return _verdict(
+                False,
+                "cannot read %s diff to secret-scan (cwd=%s) - fail-closed" % (target, cwd),            )
         hits = scan_secret_text(diff)
         if hits:
             labels = ", ".join(sorted({h[0] for h in hits}))
-            return _verdict(False, "secret in staged diff (%s) — remove it before commit/push" % labels)
+            target = "push" if push else "staged"
+            return _verdict(
+                False,
+                "secret in %s diff (%s; cwd=%s) - remove it before mutation" % (target, labels, cwd),            )
+        delivery_reason = _delivery_guard(cwd, diff, push)
+        if delivery_reason:
+            return _verdict(False, delivery_reason)
+    runtime_reason = _runtime_gate_escalation(cmd)
+    if runtime_reason:
+        return _verdict(False, "simplicio runtime gate: " + runtime_reason)
+    return _verdict(True)
+
+
     # 3) additional signal from the simplicio runtime's own risk classifier, when installed
     #    (best-effort, additive-only — see _runtime_gate_escalation).
     runtime_reason = _runtime_gate_escalation(cmd)