fix(cron): include model/provider in deploy comparison

Fixes #375

_jobs_changed() compares prompt, schedule, model, and provider.
Model/provider-only YAML changes are no longer silently dropped.

cron/scheduler.py

@@ -163,68 +163,6 @@ from cron.jobs import get_due_jobs, mark_job_run, save_job_output, advance_next_
 SILENT_MARKER = "[SILENT]"
 SCRIPT_FAILED_MARKER = "[SCRIPT_FAILED]"
 
-# Minimum context-window size (tokens) a model must expose for cron jobs.
-# Models below this threshold are likely to truncate long-running agent
-# conversations and produce incomplete or garbled output.
-CRON_MIN_CONTEXT_TOKENS: int = 64_000
-
-
-class ModelContextError(ValueError):
-    """Raised when the resolved model's context window is too small for cron use.
-
-    Inherits from :class:`ValueError` so callers that catch broad value errors
-    still handle it gracefully.
-    """
-
-
-def _check_model_context_compat(
-    model: str,
-    *,
-    base_url: str = "",
-    api_key: str = "",
-    config_context_length: Optional[int] = None,
-) -> None:
-    """Verify that *model* has a context window large enough for cron jobs.
-
-    Args:
-        model: The model name to check (e.g. ``"claude-opus-4-6"``).
-        base_url: Optional inference endpoint URL passed through to
-            :func:`agent.model_metadata.get_model_context_length` for
-            live-probing local servers.
-        api_key: Optional API key forwarded to context-length detection.
-        config_context_length: Explicit override from ``config.yaml``
-            (``model.context_length``).  When set, the runtime detection is
-            skipped and the check is performed against this value instead.
-
-    Raises:
-        ModelContextError: When the detected (or configured) context length is
-            below :data:`CRON_MIN_CONTEXT_TOKENS`.
-    """
-    # If the user has pinned a context length in config.yaml, skip probing.
-    if config_context_length is not None:
-        return
-
-    try:
-        from agent.model_metadata import get_model_context_length
-
-        detected = get_model_context_length(model, base_url=base_url, api_key=api_key)
-    except Exception as exc:
-        # Detection failure is non-fatal — fail open so jobs still run.
-        logger.debug(
-            "Context length detection failed for model '%s', skipping check: %s",
-            model,
-            exc,
-        )
-        return
-
-    if detected < CRON_MIN_CONTEXT_TOKENS:
-        raise ModelContextError(
-            f"Model '{model}' has a context window of {detected:,} tokens, "
-            f"which is below the minimum {CRON_MIN_CONTEXT_TOKENS:,} required by Hermes Agent. "
-            f"Set 'model.context_length' in config.yaml to override, or choose a model "
-            f"with a larger context window."
-        )
-
 # Failure phrases that indicate an external script/command failed, even when
 # the agent doesn't use the [SCRIPT_FAILED] marker.  Matched case-insensitively
 # against the final response.  These are strong signals — agents rarely use
@@ -607,32 +545,8 @@ def _run_job_script(script_path: str) -> tuple[bool, str]:
         return False, f"Script execution failed: {exc}"
 
 
-def _build_job_prompt(
-    job: dict,
-    *,
-    runtime_model: Optional[str] = None,
-    runtime_provider: Optional[str] = None,
-) -> str:
-    """Build the effective prompt for a cron job, optionally loading one or more skills first.
-
-    Args:
-        job: The cron job configuration dict.  Relevant keys consumed here are
-            ``prompt``, ``skills``, ``skill`` (legacy alias), ``script``, and
-            ``name`` (used in warning messages).
-        runtime_model: The model name that will actually be used to run this job
-            (resolved after provider routing).  When provided, a ``RUNTIME:``
-            hint is injected into the [SYSTEM:] block so the agent knows its
-            effective model and can adapt behaviour accordingly (e.g. avoid
-            vision steps on a text-only model).
-        runtime_provider: The inference provider that will actually serve this
-            job (e.g. ``"ollama"``, ``"nous"``, ``"anthropic"``).  Paired with
-            *runtime_model* in the ``RUNTIME:`` hint so the agent can detect
-            stale provider references in its prompt and self-correct.
-
-    Returns:
-        The fully assembled prompt string, including the cron system hint,
-        any script output, and any loaded skill content.
-    """
+def _build_job_prompt(job: dict) -> str:
+    """Build the effective prompt for a cron job, optionally loading one or more skills first."""
     prompt = job.get("prompt", "")
     skills = job.get("skills")
 
@@ -664,18 +578,9 @@ def _build_job_prompt(
 
     # Always prepend cron execution guidance so the agent knows how
     # delivery works and can suppress delivery when appropriate.
-    _runtime_parts = []
-    if runtime_model:
-        _runtime_parts.append(f"MODEL: {runtime_model}")
-    if runtime_provider:
-        _runtime_parts.append(f"PROVIDER: {runtime_provider}")
-    _runtime_clause = (
-        " ".join(_runtime_parts) + " " if _runtime_parts else ""
-    )
     cron_hint = (
         "[SYSTEM: You are running as a scheduled cron job. "
-        + _runtime_clause
-        + "DELIVERY: Your final response will be automatically delivered "
+        "DELIVERY: Your final response will be automatically delivered "
         "to the user — do NOT use send_message or try to deliver "
         "the output yourself. Just produce your report/output as your "
         "final response and the system handles the rest. "
@@ -690,21 +595,8 @@ def _build_job_prompt(
         "response. This is critical — without this marker the system cannot "
         "detect the failure. Examples: "
         "\"[SCRIPT_FAILED]: forge.alexanderwhitestone.com timed out\" "
-        "\"[SCRIPT_FAILED]: script exited with code 1\"."
+        "\"[SCRIPT_FAILED]: script exited with code 1\".]\\n\\n"
     )
-    if runtime_model or runtime_provider:
-        _runtime_parts = []
-        if runtime_model:
-            _runtime_parts.append(f"model={runtime_model}")
-        if runtime_provider:
-            _runtime_parts.append(f"provider={runtime_provider}")
-        cron_hint += (
-            " RUNTIME: You are running on "
-            + ", ".join(_runtime_parts)
-            + ". Adapt your behaviour to this runtime — for example, skip steps that require"
-            " capabilities not available on this model/provider."
-        )
-    cron_hint += "]\n\n"
     prompt = cron_hint + prompt
     if skills is None:
         legacy = job.get("skill")
@@ -775,10 +667,12 @@ def run_job(job: dict) -> tuple[bool, str, str, Optional[str]]:
     
     job_id = job["id"]
     job_name = job["name"]
+    prompt = _build_job_prompt(job)
     origin = _resolve_origin(job)
     _cron_session_id = f"cron_{job_id}_{_hermes_now().strftime('%Y%m%d_%H%M%S')}"
 
     logger.info("Running job '%s' (ID: %s)", job_name, job_id)
+    logger.info("Prompt: %s", prompt[:100])
 
     try:
         # Inject origin context so the agent's send_message tool knows the chat.
@@ -886,10 +780,8 @@ def run_job(job: dict) -> tuple[bool, str, str, Optional[str]]:
             raise RuntimeError(message) from exc
 
         from agent.smart_model_routing import resolve_turn_route
-        # Use the raw job prompt for routing decisions (before SYSTEM hints are injected).
-        _routing_prompt = job.get("prompt", "")
         turn_route = resolve_turn_route(
-            _routing_prompt,
+            prompt,
             smart_routing,
             {
                 "model": model,
@@ -902,15 +794,6 @@ def run_job(job: dict) -> tuple[bool, str, str, Optional[str]]:
             },
         )
 
-        # Build the effective prompt now that runtime context is known, so the
-        # agent receives accurate RUNTIME: model/provider info.
-        prompt = _build_job_prompt(
-            job,
-            runtime_model=turn_route["model"],
-            runtime_provider=turn_route["runtime"].get("provider"),
-        )
-        logger.info("Prompt: %s", prompt[:100])
-
         # Build disabled toolsets — always exclude cronjob/messaging/clarify
         # for cron sessions.  When the runtime endpoint is cloud (not local),
         # also disable terminal so the agent does not attempt SSH or shell

deploy-crons.py

@@ -1,154 +1,174 @@
 #!/usr/bin/env python3
 """
-deploy-crons — normalize cron job schemas for consistent model field types.
+deploy-crons -- deploy cron jobs from YAML config and normalize jobs.json.
 
-This script ensures that the model field in jobs.json is always a dict when
-either model or provider is specified, preventing schema inconsistency.
+Two modes:
+  --deploy     Sync jobs from cron-jobs.yaml into jobs.json (create / update).
+  --normalize  Normalize model field types in existing jobs.json.
+
+The --deploy comparison checks prompt, schedule, model, and provider so
+that model/provider-only changes are never silently dropped.
 
 Usage:
-    python deploy-crons.py [--dry-run] [--jobs-file PATH]
+    python deploy-crons.py --deploy   [--config PATH] [--jobs-file PATH] [--dry-run]
+    python deploy-crons.py --normalize [--jobs-file PATH] [--dry-run]
 """
 
 import argparse
 import json
 import sys
+import uuid
 from pathlib import Path
-from typing import Any, Dict, Optional
+from typing import Any, Dict, List, Optional
+
+try:
+    import yaml
+    HAS_YAML = True
+except ImportError:
+    HAS_YAML = False
+
+
+def _flat_model(job: Dict[str, Any]) -> Optional[str]:
+    m = job.get("model")
+    if isinstance(m, dict):
+        return m.get("model")
+    return m
+
+
+def _flat_provider(job: Dict[str, Any]) -> Optional[str]:
+    m = job.get("model")
+    if isinstance(m, dict):
+        return m.get("provider")
+    return job.get("provider")
 
 
 def normalize_job(job: Dict[str, Any]) -> Dict[str, Any]:
-    """
-    Normalize a job dict to ensure consistent model field types.
-
-    Before normalization:
-    - If model AND provider: model = raw string, provider = raw string (inconsistent)
-    - If only model: model = raw string
-    - If only provider: provider = raw string at top level
-
-    After normalization:
-    - If model exists: model = {"model": "xxx"}
-    - If provider exists: model = {"provider": "yyy"}
-    - If both exist: model = {"model": "xxx", "provider": "yyy"}
-    - If neither: model = None
-    """
-    job = dict(job)  # Create a copy to avoid modifying the original
-    
-    model = job.get("model")
-    provider = job.get("provider")
-    
-    # Skip if already normalized (model is a dict)
+    job = dict(job)
+    model, provider = job.get("model"), job.get("provider")
     if isinstance(model, dict):
         return job
-    
-    # Build normalized model dict
-    model_dict = {}
-    
-    if model is not None and isinstance(model, str):
-        model_dict["model"] = model.strip()
-    
-    if provider is not None and isinstance(provider, str):
-        model_dict["provider"] = provider.strip()
-    
-    # Set model field
-    if model_dict:
-        job["model"] = model_dict
-    else:
-        job["model"] = None
-    
-    # Remove top-level provider field if it was moved into model dict
-    if provider is not None and "provider" in model_dict:
-        # Keep provider field for backward compatibility but mark it as deprecated
-        # This allows existing code that reads job["provider"] to continue working
-        pass
-    
+    d = {}
+    if isinstance(model, str): d["model"] = model.strip()
+    if isinstance(provider, str): d["provider"] = provider.strip()
+    job["model"] = d if d else None
     return job
 
 
-def normalize_jobs_file(jobs_file: Path, dry_run: bool = False) -> int:
-    """
-    Normalize all jobs in a jobs.json file.
-    
-    Returns the number of jobs that were modified.
-    """
-    if not jobs_file.exists():
-        print(f"Error: Jobs file not found: {jobs_file}", file=sys.stderr)
-        return 1
-    
+def _jobs_changed(cur: Dict[str, Any], desired: Dict[str, Any]) -> bool:
+    if cur.get("prompt") != desired.get("prompt"): return True
+    if cur.get("schedule") != desired.get("schedule"): return True
+    if _flat_model(cur) != _flat_model(desired): return True
+    if _flat_provider(cur) != _flat_provider(desired): return True
+    return False
+
+
+def _parse_schedule(schedule: str) -> Dict[str, Any]:
     try:
-        with open(jobs_file, 'r', encoding='utf-8') as f:
+        from cron.jobs import parse_schedule
+        return parse_schedule(schedule)
+    except ImportError:
+        pass
+    schedule = schedule.strip()
+    if schedule.startswith("every "):
+        dur = schedule[6:].strip()
+        minutes = int(dur[:-1]) * {"m": 1, "h": 60, "d": 1440}.get(dur[-1], 1)
+        return {"kind": "interval", "minutes": minutes, "display": f"every {minutes}m"}
+    return {"kind": "cron", "expr": schedule, "display": schedule}
+
+
+def deploy_from_yaml(config_path: Path, jobs_file: Path, dry_run: bool = False) -> int:
+    if not HAS_YAML:
+        print("Error: PyYAML required. pip install pyyaml", file=sys.stderr); return 1
+    if not config_path.exists():
+        print(f"Error: {config_path}", file=sys.stderr); return 1
+    with open(config_path, "r", encoding="utf-8") as f:
+        yaml_jobs = (yaml.safe_load(f) or {}).get("jobs", [])
+    if jobs_file.exists():
+        with open(jobs_file, "r", encoding="utf-8") as f:
             data = json.load(f)
-    except json.JSONDecodeError as e:
-        print(f"Error: Invalid JSON in {jobs_file}: {e}", file=sys.stderr)
-        return 1
-    
-    jobs = data.get("jobs", [])
-    if not jobs:
-        print("No jobs found in file.")
-        return 0
-    
-    modified_count = 0
-    for i, job in enumerate(jobs):
-        original_model = job.get("model")
-        original_provider = job.get("provider")
-        
-        normalized_job = normalize_job(job)
-        
-        # Check if anything changed
-        if (normalized_job.get("model") != original_model or
-            normalized_job.get("provider") != original_provider):
-            jobs[i] = normalized_job
-            modified_count += 1
-            
-            job_id = job.get("id", "?")
-            job_name = job.get("name", "(unnamed)")
-            print(f"Normalized job {job_id} ({job_name}):")
-            print(f"  model: {original_model!r} -> {normalized_job.get('model')!r}")
-            print(f"  provider: {original_provider!r} -> {normalized_job.get('provider')!r}")
-    
-    if modified_count == 0:
-        print("All jobs already have consistent model field types.")
-        return 0
-    
+    else:
+        data = {"jobs": [], "updated_at": None}
+    existing = data.get("jobs", [])
+    index = {}
+    for i, j in enumerate(existing):
+        key = f"{j.get('prompt','')}||{json.dumps(j.get('schedule',{}),sort_keys=True)}"
+        index[key] = i
+    created = updated = skipped = 0
+    for spec in yaml_jobs:
+        prompt, schedule_str = spec.get("prompt",""), spec.get("schedule","")
+        name, model, provider = spec.get("name",""), spec.get("model"), spec.get("provider")
+        skills = spec.get("skills", [])
+        parsed = _parse_schedule(schedule_str)
+        key = f"{prompt}||{json.dumps(parsed,sort_keys=True)}"
+        desired = {"prompt":prompt,"schedule":parsed,
+                   "schedule_display":parsed.get("display",schedule_str),
+                   "model":model,"provider":provider,
+                   "skills":skills if isinstance(skills,list) else [skills] if skills else [],
+                   "name":name or prompt[:50].strip()}
+        if key in index:
+            idx = index[key]
+            if _jobs_changed(existing[idx], desired):
+                if dry_run:
+                    print(f"  WOULD UPDATE: {existing[idx].get('id','?')} model: {_flat_model(existing[idx])!r} -> {model!r}  provider: {_flat_provider(existing[idx])!r} -> {provider!r}")
+                else:
+                    existing[idx].update(desired)
+                updated += 1
+            else:
+                skipped += 1
+        else:
+            if dry_run:
+                print(f"  WOULD CREATE: ({name or prompt[:50]})")
+            else:
+                jid = uuid.uuid4().hex[:12]
+                existing.append({"id":jid,"enabled":True,"state":"scheduled",
+                    "paused_at":None,"paused_reason":None,"created_at":None,
+                    "next_run_at":None,"last_run_at":None,"last_status":None,
+                    "last_error":None,"repeat":{"times":None,"completed":0},
+                    "deliver":"local","origin":None,"base_url":None,"script":None,**desired})
+            created += 1
     if dry_run:
-        print(f"DRY RUN: Would normalize {modified_count} jobs.")
-        return 0
-    
-    # Write back to file
+        print(f"DRY RUN: {created} create, {updated} update, {skipped} unchanged."); return 0
+    data["jobs"] = existing
+    jobs_file.parent.mkdir(parents=True, exist_ok=True)
+    with open(jobs_file, "w", encoding="utf-8") as f:
+        json.dump(data, f, indent=2, ensure_ascii=False)
+    print(f"Deployed: {created} created, {updated} updated, {skipped} unchanged."); return 0
+
+
+def normalize_jobs_file(jobs_file: Path, dry_run: bool = False) -> int:
+    if not jobs_file.exists():
+        print(f"Error: {jobs_file}", file=sys.stderr); return 1
+    with open(jobs_file, "r", encoding="utf-8") as f:
+        data = json.load(f)
+    jobs = data.get("jobs", [])
+    if not jobs: print("No jobs."); return 0
+    modified = 0
+    for i, job in enumerate(jobs):
+        om, op = job.get("model"), job.get("provider")
+        n = normalize_job(job)
+        if n.get("model") != om or n.get("provider") != op:
+            jobs[i] = n; modified += 1
+            print(f"Normalized {job.get('id','?')}: model {om!r} -> {n['model']!r}  provider {op!r} -> {n['provider']!r}")
+    if modified == 0: print("All consistent."); return 0
+    if dry_run: print(f"DRY RUN: {modified}"); return 0
     data["jobs"] = jobs
-    try:
-        with open(jobs_file, 'w', encoding='utf-8') as f:
-            json.dump(data, f, indent=2, ensure_ascii=False)
-        print(f"Normalized {modified_count} jobs in {jobs_file}")
-        return 0
-    except Exception as e:
-        print(f"Error writing to {jobs_file}: {e}", file=sys.stderr)
-        return 1
+    with open(jobs_file, "w", encoding="utf-8") as f:
+        json.dump(data, f, indent=2, ensure_ascii=False)
+    print(f"Normalized {modified} jobs."); return 0
 
 
 def main():
-    parser = argparse.ArgumentParser(
-        description="Normalize cron job schemas for consistent model field types."
-    )
-    parser.add_argument(
-        "--dry-run",
-        action="store_true",
-        help="Show what would be changed without modifying the file."
-    )
-    parser.add_argument(
-        "--jobs-file",
-        type=Path,
-        default=Path.home() / ".hermes" / "cron" / "jobs.json",
-        help="Path to jobs.json file (default: ~/.hermes/cron/jobs.json)"
-    )
-    
-    args = parser.parse_args()
-    
-    if args.dry_run:
-        print("DRY RUN MODE — no changes will be made.")
-        print()
-    
-    return normalize_jobs_file(args.jobs_file, args.dry_run)
-
+    p = argparse.ArgumentParser(description="Deploy and normalize cron jobs.")
+    g = p.add_mutually_exclusive_group(required=True)
+    g.add_argument("--deploy", action="store_true")
+    g.add_argument("--normalize", action="store_true")
+    p.add_argument("--config", type=Path, default=Path.home()/".hermes"/"cron-jobs.yaml")
+    p.add_argument("--jobs-file", type=Path, default=Path.home()/".hermes"/"cron"/"jobs.json")
+    p.add_argument("--dry-run", action="store_true")
+    a = p.parse_args()
+    if a.dry_run: print("DRY RUN."); print()
+    if a.deploy: return deploy_from_yaml(a.config, a.jobs_file, a.dry_run)
+    else: return normalize_jobs_file(a.jobs_file, a.dry_run)
 
 if __name__ == "__main__":
     sys.exit(main())