agentic-dev-loop 0.15.2__py3-none-any.whl → 0.15.4__py3-none-any.whl

agentic_dev_loop/engine/brief_coverage.py

@@ -11,11 +11,26 @@ and file-agnostic.
 from __future__ import annotations
 
 import re
+import unicodedata
+
+
+_GLYPH_FOLD = str.maketrans({
+    "\u2010": "-", "\u2011": "-", "\u2012": "-",
+    "\u2013": "-", "\u2014": "-", "\u2015": "-",
+    "\u2018": "'", "\u2019": "'",
+    "\u201c": '"', "\u201d": '"',
+})
+
+
+def _fold_glyphs(text: str) -> str:
+    """Fold compatibility glyphs and collapse whitespace, preserving case."""
+    folded = unicodedata.normalize("NFKC", text).translate(_GLYPH_FOLD)
+    return " ".join(folded.split())
 
 
 def _normalize(text: str) -> str:
-    """Case-fold and collapse whitespace - the matching normal form."""
-    return " ".join(text.split()).casefold()
+    """Case-folded, glyph-folded, whitespace-collapsed matching form."""
+    return _fold_glyphs(text).casefold()
 
 
 def _normalize_line(line: str) -> str:
@@ -23,7 +38,7 @@ def _normalize_line(line: str) -> str:
 
     A compliant digest writes anchors as headings on their own line, wrapped
     in ``**`` and (for the four section headings) prefixed with ``##``. List
-    bullets are tolerated; the whole-line comparison is what makes a mid-line
+    bullets are tolerated; the line-prefix comparison is what makes a mid-line
     prose mention fail."""
     stripped = line.strip().replace("**", "")
     stripped = re.sub(r"^#+\s*", "", stripped)
@@ -67,19 +82,27 @@ def missing_anchors(brief_text: str, anchors: list[str],
     """Return every required item the captured brief fails to cover.
 
     Required section headings (from the configured ``brief_inputs.sections``)
-    and anchors are matched line-anchored (case-insensitive,
-    whitespace-collapsed, ``**`` decoration stripped). Env facts are matched
-    verbatim, exactly as handed to the brief agent. An empty list means the
+    and anchors are matched as line prefixes (case-insensitive,
+    glyph-folded, whitespace-collapsed, decoration stripped): an item is
+    covered when it starts a line, so content may follow on that line. A
+    mid-line mention does not count. Env facts are matched as glyph-folded,
+    whitespace-collapsed, case-sensitive substrings. An empty list means the
     digest passes coverage."""
     missing: list[str] = []
-    lines = {_normalize_line(line) for line in brief_text.splitlines()}
+    lines = [_normalize_line(line) for line in brief_text.splitlines()]
+    lines = [line for line in lines if line]
+
+    def covered(item: str) -> bool:
+        return any(line.startswith(item) for line in lines)
+
     for section in sections:
-        if _normalize(section) not in lines:
+        if not covered(_normalize(section)):
             missing.append(section)
     for anchor in anchors:
-        if anchor not in lines:
+        if not covered(anchor):
             missing.append(anchor)
+    folded_brief = _fold_glyphs(brief_text)
     for fact in env_facts:
-        if fact not in brief_text:
+        if _fold_glyphs(fact) not in folded_brief:
             missing.append(fact)
     return missing

agentic_dev_loop/engine/brief_prompt.md

@@ -26,21 +26,24 @@ a compact brief that the dev and review agents reuse on every attempt.
   attempt. Do not invent them and do not re-derive them.
 - The repo-rules digest is structural: one line per top-level section of each
   repo-rule input. Markdown anchors are the `##` headings: keep each heading
-  verbatim, then compress that section's rules into the one line that follows
-  it. YAML anchors are each `- name:` entry under `categories:` plus the
+  verbatim at the start of its own line, then compress that section's rules
+  into the text that follows on that line (after a colon) or on the next
+  line. YAML anchors are each `- name:` entry under `categories:` plus the
   `enforcement_levels` preamble: keep each name verbatim. No top-level section
   or category may be dropped; merging is allowed only between a section and
-  its own subsections. The anchor line holds the heading and nothing else; the
-  compressed rules go on the following line:
+  its own subsections:
 
   - general_principles
     Simplest thing that works; defer decisions to the last responsible moment.
 
-  A line like `general_principles: Simplest thing that works; ...` does not
-  satisfy coverage; the heading must be the only text on its line.
+  A line like `general_principles: Simplest thing that works; ...` passes too:
+  the heading must start its line.
 - Keep the environment-facts entries verbatim: each fact and the behavior it
   implies travel together.
-- ASCII only. Do not use em dash characters.
+- ASCII only, except that a required heading, env fact, or configured section
+  heading must be reproduced character-exact: when such an item contains a
+  non-ASCII character, keep that exact character, never transliterate it. Do
+  not introduce em dashes or other non-ASCII characters of your own.
 - If the task message carries a `## Coverage retry` section, regenerate the
   brief so every listed item appears, verbatim where the contract requires
   verbatim. No exploration: the listed items are the only correction signal.

agentic_dev_loop/engine/brief_prompt.py

@@ -15,7 +15,7 @@ from __future__ import annotations
 from importlib.resources import files
 
 BRIEF_PROMPT_RESOURCE = "brief_prompt.md"
-BRIEF_PROMPT_VERSION = "1"
+BRIEF_PROMPT_VERSION = "2"
 
 
 def load_template() -> str:
@@ -36,9 +36,9 @@ def _output_block(sections: list[str]) -> str:
         "compressed content for that section (write `None` when a section has "
         "nothing):\n\n"
         f"{headings}\n\n"
-        f"The section headings must appear exactly as written above: {names}. "
-        "Do not shorten or rename them, and do not merge a section with "
-        "another section."
+        "The section headings must appear at the start of their own lines, as "
+        f"written above: {names}. Do not shorten or rename them, and do not "
+        "merge a section with another section."
     )
 
 

agentic_dev_loop/engine/onboarding/templates/dev.agent.md

@@ -10,10 +10,11 @@ opinionated LangGraph + OpenHands SDK pipeline engine written in Python.
 ## WORKFLOW
 
 1. Read the plan documents and README.md in the context below.
-2. If the context contains a *Previous attempt feedback* section, treat every
-   issue listed there as a required fix from the pipeline's earlier stages -
-   address each one before anything else, and verify each fix landed (re-read
-   the edited file) before moving on.
+2. If the context contains a *Previous attempt feedback* section or a
+   *Mechanical gate output* section, treat every issue listed there as a
+   required fix from the pipeline's earlier stages - address each one before
+   anything else, and verify each fix landed (re-read the edited file) before
+   moving on. A mechanical entry names the attempt that produced it.
 3. Check for prior-attempt residue before implementing. The pipeline syncs
    every lane's edits back into this repo's working tree after each terminal
    action, so an earlier attempt that died at a later gate can leave the full
@@ -66,10 +67,23 @@ opinionated LangGraph + OpenHands SDK pipeline engine written in Python.
   exit code alone.
 - You do not need to prove anything here: fill in the verification matrix,
   write your final response, and stop.
-- If the gate fails, you are retried with the failing output in your notes
-  under *Previous attempt feedback*. Find the root cause - a regression in
-  your change, a stale test pinning old behavior, or a genuine library bug -
-  and fix it, never the symptom.
+- If the gate fails, you are retried with the failing output under
+  *Mechanical gate output* and the exit code under *Verified state* (command
+  gates do not write *Previous attempt feedback*). Find the root cause - a
+  regression in your change, a stale test pinning old behavior, or a genuine
+  library bug - and fix it, never the symptom.
+
+## AUDIT GATE
+
+- The restore step re-runs the install and a dependency audit on every pass.
+  A failing audit means a pin in the dependency manifest carries known CVEs;
+  it is a required fix, never "pre-existing" or "unrelated".
+- Fix it with a manifest edit, never a package-manager command (agent steps
+  are offline). Follow the floor-pin pattern: pin the vulnerable resolved
+  package to a CVE-free version, or pin the intermediate package to the last
+  release that admits it, so a fresh resolve lands on the fixed version.
+- Restore re-audits the bumped pins on the next pass, so the loop
+  self-corrects; report the audit as fixed only after that gate passes.
 
 ## TEST-FIRST
 
@@ -81,7 +95,8 @@ opinionated LangGraph + OpenHands SDK pipeline engine written in Python.
 
 - The moment your final response is written AND the full suite has passed
   locally (`.venv-sandbox/bin/python -m pytest tests/ -q` exits 0) AND every
-  *Previous attempt feedback* item is addressed, the step is COMPLETE. Stop.
+  *Previous attempt feedback* and *Mechanical gate output* item is addressed,
+  the step is COMPLETE. Stop.
 - Do not re-read files, re-run checks, or `cat`/`tail`/`grep` files you just
   wrote to confirm what already passed. A green suite is a green suite:
   re-running it buys nothing, since the mechanical test gate re-runs the

agentic_dev_loop/engine/stages.py

@@ -446,7 +446,10 @@ def build_context(runtime: Runtime, docs: dict[str, str], subject: str,
     own mechanical section (process fact, never an agent claim), and only
     engine-verified facts (command exit codes, brief coverage, commit outcome)
     appear in the verified section. Nothing in the claims section is ever
-    routed on."""
+    routed on. The mechanical heading binds every failure it carries as a
+    required fix: command gates (test, restore) never write the notes channel,
+    so without that binding a retried dev agent has no required-fix trigger
+    (plan 065)."""
     parts = [f"You are working on {subject}.\n", guide_section()]
     for name, content in docs.items():
         parts.append(f"## {name}\n\n{content}")
@@ -470,7 +473,8 @@ def build_context(runtime: Runtime, docs: dict[str, str], subject: str,
         parts.extend(claims)
     if mechanical_notes:
         parts.append("## Mechanical gate output "
-                     "(pipeline-captured process output, not agent claims)")
+                     "(pipeline-captured process output, not agent claims; "
+                     "each failure here is a required fix)")
         parts.extend(mechanical_notes)
     verified_lines = [f"- {name}: {fact}" for name, fact in (verified or {}).items()]
     parts.append(
@@ -678,16 +682,11 @@ def build_brief_context(runtime: Runtime, step: StepConfig,
         parts.append(
             "## Coverage retry\n\n"
             "The captured brief is missing the following required items. "
-            "Regenerate the digest so every listed item occupies its own "
-            "line, exactly as listed: the item alone on the line, no "
-            "trailing colon and no content. Put the compressed content on "
-            "the following line:\n\n"
+            "Regenerate the digest so every listed item starts its own line; "
+            "compressed content may follow on that line or the next line.\n\n"
             "- general_principles\n"
             "  Simplest thing that works; defer decisions to the last "
             "responsible moment.\n\n"
-            "A line like `general_principles: Simplest thing that works; "
-            "...` does not satisfy coverage; the item must be the only "
-            "text on its line:\n\n"
             + coverage_note)
     return "\n\n".join(parts)
 

{agentic_dev_loop-0.15.2.dist-info → agentic_dev_loop-0.15.4.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agentic-dev-loop
-Version: 0.15.2
+Version: 0.15.4
 Summary: Turn an LLM into a gated, sandboxed, resumable construction loop over your codebase.
 Author: Luke Wilson
 License-Expression: Apache-2.0
@@ -29,6 +29,8 @@ Requires-Dist: pydantic>=2.0
 Requires-Dist: PyYAML>=6.0
 Requires-Dist: openhands-sdk~=1.17.0
 Requires-Dist: openhands-tools~=1.17.0
+Requires-Dist: browser-use==0.11.13
+Requires-Dist: pillow>=13.0.0
 Requires-Dist: agent-client-protocol<1.0.0
 Requires-Dist: langgraph>=1.2
 Requires-Dist: langgraph-checkpoint-sqlite>=3.0
@@ -560,10 +562,11 @@ Conventions that make a good agent file:
 - **dev enumerates its impossible set and stops at it.** List the commands
   the sandbox cannot run (offline package fetches, browser downloads), one
   attempt per command - keeps the productive loop and kills doomed retries.
-- **Failure output returns as *Previous attempt feedback*.** Steps route
-  `on_fail: { goto: dev, retry: true }`, so gate failure output lands in the
-  dev agent's next context; the dev prompt makes every listed issue a
-  required fix.
+- **Failure output is a required fix in dev's next context.** Agent verdict
+  steps land their failing output under *Previous attempt feedback*; command
+  gates (test, restore) land it under *Mechanical gate output*, with the exit
+  code under *Verified state*. Steps route `on_fail: { goto: dev, retry: true }`,
+  and the dev prompt makes every listed failure a required fix.
 - **No git mutations, ever.** The `commit` step is the single commit point.
 
 ### The pipeline file

{agentic_dev_loop-0.15.2.dist-info → agentic_dev_loop-0.15.4.dist-info}/RECORD

@@ -15,9 +15,9 @@ agentic_dev_loop/engine/agent_error_guide.md,sha256=PL8WbDWs6W0TUf0AMz1a6Dh0YbfV
 agentic_dev_loop/engine/agent_error_guide.py,sha256=RI64yC9kzlYabKSj78F2gY_qmYjnZlCi-3ImQBZEJu4,798
 agentic_dev_loop/engine/agents.py,sha256=I5WbVB-mCteY-JOqPY2tGnlqP5WSpCb7SNqYKegU_jY,13707
 agentic_dev_loop/engine/allowlist.py,sha256=IrtWNyNNojMXGPDUJCLLcdlWm4Fn6yNwD4VvlMme6wM,2453
-agentic_dev_loop/engine/brief_coverage.py,sha256=7wL8aP1_dBl62hh0IyZ60bpN4-hfOOsQDKURmeLi8uo,3356
-agentic_dev_loop/engine/brief_prompt.md,sha256=0MmA4xNdeJCtjoZxjE3iNm5J_k_uCqrajOXxR5hyzYM,2464
-agentic_dev_loop/engine/brief_prompt.py,sha256=nHkGFdyWkZcZyDEZHAsLIO3pUzpw2nsCHk8QYwqH8J4,1904
+agentic_dev_loop/engine/brief_coverage.py,sha256=E76bIX6CVWBd41-vgAeU5DYBLwybmpBFMxX6AhKFe-4,4150
+agentic_dev_loop/engine/brief_prompt.md,sha256=s69XFuxN94bzDo_YADE0sT3sYI9bYIpzrkbByyIq2PE,2661
+agentic_dev_loop/engine/brief_prompt.py,sha256=NlGUnBI2gk37YisxnVr7P0WcuCq1bMM7zcROQV76UKI,1929
 agentic_dev_loop/engine/cli.py,sha256=1W20QTPEzyJCnGnE58_pa4TiZeJPAb4YX38WR3odNW8,20276
 agentic_dev_loop/engine/completion.py,sha256=JguMetYsXKm02Rmq0o360oNt4ElXlKXzFILhv-P7xg8,1939
 agentic_dev_loop/engine/containers.py,sha256=r2iVEzEkF422OTZEUFnROT92b2HLRfrMtvfDbYPGcvY,3541
@@ -41,7 +41,7 @@ agentic_dev_loop/engine/registry.py,sha256=qepez9iU-dBE-yG67qI7Q7Ol89NsERHy34eoQ
 agentic_dev_loop/engine/run_status.py,sha256=d_jw_3hT1oEqG8j5hDfq7XT71hZyoKhgJDjNYd88ZIE,12144
 agentic_dev_loop/engine/run_status_command.md,sha256=PhusrtsgcCGHTIEb7xQvcK46i-NdZIHi_zEG-MbsGx0,1357
 agentic_dev_loop/engine/secrets.py,sha256=xdxw-JwmgDkiLdJkBchym9UADZiQ5M4NagmdJhToQYI,8386
-agentic_dev_loop/engine/stages.py,sha256=_g5K7jg53UK2i1KeE2uM8dau-siyILaWK-vhrE2nvGY,63319
+agentic_dev_loop/engine/stages.py,sha256=8oi1RKfFQznbflJrW5GXwd8AQOL47qzDlvVvZgX-cuE,63324
 agentic_dev_loop/engine/step_workspace.py,sha256=YyaZWk8ozc_v7AFBt63QAn4VFL-a1wvhzNvnbascYYw,7937
 agentic_dev_loop/engine/sync_back.py,sha256=Mc6YAAhc_Iydz-7Ht3OkyGekaa6s1HitCsOlIepDfOc,13942
 agentic_dev_loop/engine/extensions/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -55,7 +55,7 @@ agentic_dev_loop/engine/onboarding/templates.py,sha256=4vlu4SXY2I5BJboHnmdCHxW-l
 agentic_dev_loop/engine/onboarding/upgrade.py,sha256=5WiqLlef30yg78XRYZQWyHcYIIsCNRoxbv_UL2bnAxY,2701
 agentic_dev_loop/engine/onboarding/recipes/dotnet-node.yaml,sha256=Uqq__RbPbZ3NuNtW04-CAdY-mARP3oQHzt2byTZMz_g,3743
 agentic_dev_loop/engine/onboarding/recipes/python.yaml,sha256=42nvggXaQ4rrPAlNjrYnLeD6czuwchYVRnrFziefRKA,2491
-agentic_dev_loop/engine/onboarding/templates/dev.agent.md,sha256=GdbENdOrWSJ8FKWqf97Z9JsjnEZuC9p0RHTNR6Cmuk8,6460
+agentic_dev_loop/engine/onboarding/templates/dev.agent.md,sha256=37sjR-WXdTZWQ6baPN4Tn_DjDWWPuTurg3-fxZm20rU,7339
 agentic_dev_loop/engine/onboarding/templates/env.example,sha256=o_BzGtkXsRKjiKeWpsz4mVqPnHstETBF-PsvwJawN8g,532
 agentic_dev_loop/engine/onboarding/templates/qa.agent.md,sha256=uA_7xJZKsW3L5UFvvR-u2lHfiXcRgvhnTdFaUjmAVZE,10389
 agentic_dev_loop/engine/onboarding/templates/review.agent.md,sha256=_88SgIlXLSDowpRloZl985zq1v4O9b7NaUXEt1-9u-U,2069
@@ -70,9 +70,9 @@ agentic_dev_loop/viz/layout.py,sha256=yXBT96rRGgrNm1aumZ6A8SZCeqiT6akknc1GMJn2Hh
 agentic_dev_loop/viz/page.py,sha256=JkiJS6ixCfRLG6EVS2Av2aEU8YfSF20QViSvzYdx1sk,1278233
 agentic_dev_loop/viz/runs_view.py,sha256=EaXHvH1y7TL4fcXzj7X913ckdTPv6R21HD1ivGwKE0c,6780
 agentic_dev_loop/viz/server.py,sha256=X65h6gyx8NCtl32_NGN_01zsTjoF3Nzd5Y3NdfxjhjY,23371
-agentic_dev_loop-0.15.2.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-agentic_dev_loop-0.15.2.dist-info/METADATA,sha256=mKBN26e7EpAdaWoKqaRtnorKJauckML4-Hhm4eHKmaU,33600
-agentic_dev_loop-0.15.2.dist-info/WHEEL,sha256=mOdXR9C3wMVov7cSY8PY-XCzo3SmmWCHvGQr5csJoao,91
-agentic_dev_loop-0.15.2.dist-info/entry_points.txt,sha256=Us59oHX6NBy-7ZN3LW5w3HAjnvOPJBfgNiv6ftnmHWw,70
-agentic_dev_loop-0.15.2.dist-info/top_level.txt,sha256=n1ktP1gXxGN6DJAnACR0mt276GRJB9PVDezf9DVPwaw,17
-agentic_dev_loop-0.15.2.dist-info/RECORD,,
+agentic_dev_loop-0.15.4.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+agentic_dev_loop-0.15.4.dist-info/METADATA,sha256=0wuRMEsNVFQ_DKhHGvu1_v7fBzDNthBz-4Hut37P7bA,33807
+agentic_dev_loop-0.15.4.dist-info/WHEEL,sha256=mOdXR9C3wMVov7cSY8PY-XCzo3SmmWCHvGQr5csJoao,91
+agentic_dev_loop-0.15.4.dist-info/entry_points.txt,sha256=Us59oHX6NBy-7ZN3LW5w3HAjnvOPJBfgNiv6ftnmHWw,70
+agentic_dev_loop-0.15.4.dist-info/top_level.txt,sha256=n1ktP1gXxGN6DJAnACR0mt276GRJB9PVDezf9DVPwaw,17
+agentic_dev_loop-0.15.4.dist-info/RECORD,,