Eliminate conversational prompting in taskflow execution

Add explicit execution directives to personality and all task prompts:
- 'Execute all steps above, then stop. Do not ask what to do next.'
- Personality: 'Do not offer to do additional work. Do not say if
  you'd like or shall I. Complete every step, then stop.'

Prevents the model from pausing mid-task to ask for user direction
instead of executing the defined workflow autonomously.

Author: anticomputer

src/seclab_taskflows/configs/model_config_pvr_triage.yaml

@@ -11,7 +11,7 @@ seclab-taskflow-agent:
 
 models:
   # Primary model for code analysis and triage reasoning
-  triage: claude-opus-4.6-1m
+  triage: claude-opus-4.6
   # Lighter model for structured data extraction tasks
   extraction: gpt-5-mini
 

src/seclab_taskflows/personalities/pvr_analyst.yaml

@@ -21,9 +21,17 @@ personality: |
   - Flag low-quality ("AI slop") reports: vague claims, wrong file paths, non-working PoC,
     incorrect function signatures, or descriptions that don't match the actual code.
 
+  Execution rules:
+  - Execute each task fully and autonomously. Do not ask for permission or confirmation.
+  - Do not offer to do additional work. Do not say "if you'd like" or "shall I".
+  - Complete every step described in the task prompt, then stop.
+  - When storing results in memcache, proceed immediately to the next step.
+  - Output only findings and results, not suggestions for next steps.
+
 task: |
   Analyze the provided vulnerability report and verify claims against the actual source code.
   Produce factual, evidence-based findings. Never guess or assume.
+  Execute all steps described in the user prompt without asking for direction.
 
 toolboxes:
   - seclab_taskflows.toolboxes.pvr_ghsa

src/seclab_taskflows/taskflows/pvr_triage/pvr_triage.yaml

@@ -95,6 +95,8 @@ taskflow:
         is returned (non-empty), store it under memcache key "security_policy".
         If no policy is found, store an empty string.
 
+        Execute all steps above, then stop. Do not ask what to do next.
+
   # -------------------------------------------------------------------------
   # Task 3: Quick Quality Gate
   # -------------------------------------------------------------------------
@@ -208,6 +210,8 @@ taskflow:
         but do NOT auto-set fast_close. Duplicates still require human judgment.
         The report will surface the duplicate info for the maintainer to decide.
 
+        Execute all steps above, then stop. Do not ask what to do next.
+
   # -------------------------------------------------------------------------
   # Task 4: Verify vulnerability in source code
   # -------------------------------------------------------------------------
@@ -282,6 +286,8 @@ taskflow:
           - patch_notes: description of HEAD state vs claimed version (or null)
           - notes: any additional observations
 
+        Execute all steps above, then stop. Do not ask what to do next.
+
   # -------------------------------------------------------------------------
   # Task 4b: Container-based validation (optional)
   # -------------------------------------------------------------------------
@@ -530,6 +536,8 @@ taskflow:
             "quality": "High" | "Medium" | "Low"
           }
 
+        Execute all steps above, then stop. Do not ask what to do next.
+
   # -------------------------------------------------------------------------
   # Task 6: Save report to disk and print path
   # -------------------------------------------------------------------------
@@ -593,6 +601,8 @@ taskflow:
 
         Store under memcache key "response_draft".
 
+        Execute all steps above, then stop. Do not ask what to do next.
+
   # -------------------------------------------------------------------------
   # Task 8: Update Reporter Reputation + Save Response Draft
   # -------------------------------------------------------------------------

src/seclab_taskflows/taskflows/pvr_triage/pvr_triage_batch.yaml

@@ -52,9 +52,11 @@ taskflow:
 
         Store the full JSON list under memcache key "pvr_queue".
 
-        Print: "Found N triage advisories for {{ globals.repo }}." where N is the count.
+        Print: "Found N {{ globals.state }} advisories for {{ globals.repo }}." where N is the count.
 
-        If no advisories are found, print "No triage advisories found." and stop.
+        If no advisories are found, print "No {{ globals.state }} advisories found." and stop.
+
+        Execute all steps above, then stop. Do not ask what to do next.
 
   # -------------------------------------------------------------------------
   # Task 2: Score each advisory and detect duplicates
@@ -138,6 +140,8 @@ taskflow:
         Store scored_queue under memcache key "scored_queue".
         Store skipped_count under memcache key "skipped_count".
 
+        Execute all steps above, then stop. Do not ask what to do next.
+
   # -------------------------------------------------------------------------
   # Task 3: Generate and save ranked queue report
   # -------------------------------------------------------------------------
@@ -206,3 +210,5 @@ taskflow:
 
         Print: "Batch queue report saved to: <path returned by save_triage_report>"
         Then print the full report.
+
+        Execute all steps above, then stop. Do not ask what to do next.