fix: Resolve all lint errors blocking consumer repo sync

scripts/langchain/capability_check.py

@@ -3,7 +3,8 @@
 Capability classification for agent issue intake.
 
 Run with:
-    python scripts/langchain/capability_check.py --tasks-file tasks.md --acceptance-file acceptance.md
+    python scripts/langchain/capability_check.py \
+        --tasks-file tasks.md --acceptance-file acceptance.md
 """
 
 from __future__ import annotations
@@ -240,7 +241,9 @@ def _fallback_classify(
                 {
                     "task": task,
                     "reason": "Requires external service credentials or configuration",
-                    "suggested_action": "Provide credentials or have a human set up the external service.",
+                    "suggested_action": (
+                        "Provide credentials or have a human set up " "the external service."
+                    ),
                 }
             )
             human_actions.append(f"External dependency setup required: {task}")

scripts/langchain/followup_issue_generator.py

@@ -69,7 +69,8 @@
 Analyze what went wrong and what SPECIFICALLY needs to change. Focus on:
 
 1. **Which original acceptance criteria are actually still unmet?**
-   - Don't assume all criteria need rework. Only include criteria that the verification shows are genuinely incomplete.
+   - Don't assume all criteria need rework. Only include criteria that the
+     verification shows are genuinely incomplete.
    - Rewrite criteria that were unclear or unmeasurable as clear, testable statements.
 
 2. **What concrete code changes are needed?**
@@ -79,7 +80,8 @@
 
 3. **Did previous iterations reveal blockers the next agent should avoid?**
    - Only include if there's specific, detailed information about what didn't work and why
-   - "Iteration 3 failed" is NOT useful. "Iteration 3 attempted X approach but failed because Y, so try Z instead" IS useful
+   - "Iteration 3 failed" is NOT useful. "Iteration 3 attempted X approach but
+     failed because Y, so try Z instead" IS useful
 
 Output JSON:
 {{
@@ -115,11 +117,13 @@
 - Sized appropriately: not too big ("fix everything") or too small ("add a comma")
 
 **Tasks MUST NOT be:**
-- Verification concerns restated as tasks (e.g., "The safety rules section is incomplete" is NOT a task)
+- Verification concerns restated as tasks (e.g., "The safety rules section
+  is incomplete" is NOT a task)
 - Original acceptance criteria restated as tasks
 - Vague actions like "improve", "ensure", "address concerns about"
 
-**Deferred items:** Anything requiring credentials, external APIs, manual testing, or human decisions
+**Deferred items:** Anything requiring credentials, external APIs, manual
+testing, or human decisions
 
 Output JSON:
 {{
@@ -217,11 +221,13 @@
 ```
 
 ## Critical Rules
-1. Do NOT include "Remaining Unchecked Items" or "Iteration Details" sections unless they contain specific, useful failure context
+1. Do NOT include "Remaining Unchecked Items" or "Iteration Details" sections
+   unless they contain specific, useful failure context
 2. Tasks should be concrete actions, not verification concerns restated
 3. Acceptance criteria must be testable (not "all concerns addressed")
 4. Keep the main body focused - hide background/history in the collapsible section
-5. Do NOT include the entire analysis object - only include specific failure contexts from `blockers_to_avoid`
+5. Do NOT include the entire analysis object - only include specific failure
+   contexts from `blockers_to_avoid`
 
 Output the complete markdown issue body.
 """.strip()
@@ -548,7 +554,10 @@ def _prepare_iteration_details(codex_log: str) -> str:
                 useful_lines.append(context_block)
 
     if not useful_lines:
-        return "Previous iterations completed without recorded failures. No specific blockers to avoid."
+        return (
+            "Previous iterations completed without recorded failures. "
+            "No specific blockers to avoid."
+        )
 
     # Deduplicate and limit length
     unique_blocks = list(dict.fromkeys(useful_lines))[:5]  # Max 5 failure contexts

scripts/langchain/issue_optimizer.py

@@ -792,7 +792,8 @@ def apply_suggestions(
                     # If GitHub Models hit token limit, retry with OpenAI API
                     if _is_token_limit_error(e) and provider == "github-models":
                         print(
-                            "GitHub Models token limit hit in apply_suggestions, retrying with OpenAI API...",
+                            "GitHub Models token limit hit in apply_suggestions, "
+                            "retrying with OpenAI API...",
                             file=sys.stderr,
                         )
                         openai_client_info = _get_llm_client(force_openai=True)

scripts/langchain/pr_verifier.py

@@ -21,7 +21,8 @@
 from pydantic import BaseModel, Field, ValidationError
 
 PR_EVALUATION_PROMPT = """
-You are reviewing a **merged** pull request to evaluate whether the code changes meet the documented acceptance criteria.
+You are reviewing a **merged** pull request to evaluate whether the code
+changes meet the documented acceptance criteria.
 
 **IMPORTANT: This verification runs AFTER the PR has been merged.** Therefore:
 - Do NOT evaluate CI status, workflow runs, or pending checks - these are irrelevant post-merge
@@ -484,8 +485,10 @@ def evaluate_pr(
     Args:
         context: The PR context markdown (issue body, PR description, etc.)
         diff: Optional PR diff or summary
-        model: Optional model name (e.g., 'gpt-4o', 'gpt-5.2', 'o1-mini'). Uses default if not specified.
-        provider: Optional provider ('openai' or 'github-models'). Auto-selects if not specified.
+        model: Optional model name (e.g., 'gpt-4o', 'gpt-5.2', 'o1-mini').
+            Uses default if not specified.
+        provider: Optional provider ('openai' or 'github-models').
+            Auto-selects if not specified.
 
     Returns:
         EvaluationResult with verdict, scores, and concerns.
@@ -525,7 +528,8 @@ def evaluate_pr(
                     return result
                 except Exception as fallback_exc:
                     return _fallback_evaluation(
-                        f"Primary ({provider_name}): {exc}; Fallback ({fallback_provider_name}): {fallback_exc}"
+                        f"Primary ({provider_name}): {exc}; "
+                        f"Fallback ({fallback_provider_name}): {fallback_exc}"
                     )
         return _fallback_evaluation(f"LLM invocation failed: {exc}")
 
@@ -622,9 +626,8 @@ def format_comparison_report(results: list[EvaluationResult]) -> str:
         summary_source = result.summary or result.raw_content or ""
         summary = _compact_text(summary_source, limit=200) if summary_source else "N/A"
         model_name = result.model or "N/A"
-        lines.append(
-            f"| {labels[index]} | {model_name} | {result.verdict} | {_format_confidence(result.confidence)} | {summary} |"
-        )
+        conf = _format_confidence(result.confidence)
+        lines.append(f"| {labels[index]} | {model_name} | {result.verdict} | {conf} | {summary} |")
     lines.append("")
 
     # Add expandable full details for each provider
@@ -748,7 +751,10 @@ def main() -> None:
     parser.add_argument(
         "--provider",
         choices=["openai", "github-models"],
-        help="LLM provider: 'openai' (requires OPENAI_API_KEY) or 'github-models' (uses GITHUB_TOKEN).",
+        help=(
+            "LLM provider: 'openai' (requires OPENAI_API_KEY) or "
+            "'github-models' (uses GITHUB_TOKEN)."
+        ),
     )
     parser.add_argument(
         "--model2",