Files
hermes-agent/hermes_cli/curator.py
Teknium 97acd66b4c fix(curator): authoritative absorbed_into on delete + restore cron skill links on rollback (#18671) (#18731)
* fix(curator): authoritative absorbed_into declarations on skill delete

Closes #18671. The classification pipeline that feeds cron-ref rewriting
used to infer consolidation vs pruning from two brittle signals: the
curator model's post-hoc YAML summary block, and a substring heuristic
scanning other tool calls for the removed skill's name. Both miss in
real consolidations — the model forgets the YAML under reasoning
pressure, and the heuristic misses when the umbrella's patch content
describes the absorbed behavior abstractly instead of naming the old
slug. When both miss, the skill falls through to 'no-evidence fallback'
pruned, and #18253's cron rewriter drops the cron ref entirely instead
of mapping it to the umbrella. Same observable symptom as pre-#18253:
'Skill(s) not found and skipped' at the next cron run.

The fix makes the model declare intent at the moment of deletion.
skill_manage(action='delete') now accepts absorbed_into:
  - absorbed_into='<umbrella>'  -> consolidated, target must exist on disk
  - absorbed_into=''            -> explicit prune, no forwarding target
  - missing                     -> legacy path, falls through to heuristic/YAML

The curator reconciler reads these declarations off llm_meta.tool_calls
BEFORE either the YAML block or the substring heuristic. Declaration
wins. Fallback logic stays intact for backward compat with any caller
(human or older curator conversation) that doesn't populate the arg.

Changes
- tools/skill_manager_tool.py: add absorbed_into param to skill_manage
  + _delete_skill. Validate target exists when non-empty. Reject
  absorbed_into=<self>. Wire through dispatcher + registry + schema.
- agent/curator.py: new _extract_absorbed_into_declarations() walks
  tool calls for skill_manage(delete) with the arg. _reconcile_classification
  accepts absorbed_declarations= and treats them as authoritative. Curator
  prompt updated to require the arg on every delete.
- Tests: 7 new skill_manager tests covering the tool contract (valid
  target, empty string, nonexistent target, self-reference, whitespace,
  backward compat, dispatcher plumbing). 11 new curator tests covering
  the extractor + authoritative reconciler path + mixed-legacy-and-
  declared runs.

Validation
- 307/307 targeted tests pass (curator + cron + skill_manager suites).
- E2E #18671 repro: 3 narrow skills, 1 umbrella, cron job referencing
  all 3. Model emits NO YAML block. Heuristic misses (patch prose
  doesn't name old slugs). Delete calls carry absorbed_into. Result:
  both PR skills correctly classified 'consolidated' + cron rewritten
  ['pr-review-format', 'pr-review-checklist', 'stale-junk'] ->
  ['hermes-agent-dev']; stale-junk pruned via absorbed_into=''.
- E2E backward-compat: delete without absorbed_into, model emits YAML
  -> routed via existing 'model' source, cron still rewritten correctly.

* feat(curator): capture + restore cron skill links across snapshot/rollback

Before this, rolling back a curator run restored the skills tree but cron
jobs still pointed at the umbrella skills the curator had rewritten them
to. The user would see their old narrow skills back on disk but their
cron jobs still configured with the merged umbrella — not actually 'back
to how it was'.

Snapshot side: snapshot_skills() now captures ~/.hermes/cron/jobs.json
alongside the skills tarball, as cron-jobs.json. The manifest gets a new
'cron_jobs' block with {backed_up, jobs_count} so rollback (and the CLI
confirm dialog) can surface what's in the snapshot. If jobs.json is
missing/unreadable/malformed, snapshot proceeds without cron data — the
skills backup is the core guarantee; cron is additive.

Rollback side: after the skills extract succeeds, the new
_restore_cron_skill_links() reconciles the backed-up jobs into the live
jobs.json SURGICALLY. Only 'skills' and 'skill' fields are restored, and
only on jobs matched by id. Everything else about a cron job — schedule,
last_run_at, next_run_at, enabled, prompt, workdir, hooks — is live
state the user or scheduler has modified since the snapshot; overwriting
it would regress unrelated activity.

Reconciliation rules:
- Job in backup AND live, skills differ  → skills restored.
- Job in backup AND live, skills match   → no-op.
- Job in backup, NOT in live             → skipped (user deleted it
                                              after snapshot; their choice
                                              is later than the snapshot).
- Job in live, NOT in backup             → untouched (user created it
                                              after snapshot).
- Snapshot missing cron-jobs.json at all → rollback still succeeds,
                                              reports 'not captured'
                                              (older pre-feature snapshots
                                              keep working).

Writes go through cron.jobs.save_jobs under the same _jobs_file_lock the
scheduler uses, so rollback doesn't race tick().

Also:
- hermes_cli/curator.py: rollback confirm dialog now shows
  'cron jobs: N (will be restored for skill-link fields only)' when the
  snapshot has cron data, or 'not in snapshot (<reason>)' otherwise.
- rollback()'s message string includes a 'cron links: ...' clause
  summarizing the reconciliation outcome.

Tests
- 9 new cases: snapshot-with-cron, snapshot-without-cron, malformed-json
  captured-as-raw, full rollback-restores-skills-and-cron, rollback
  touches only skill fields, rollback skips user-deleted jobs, rollback
  leaves user-created jobs untouched, rollback still works with
  pre-feature snapshot that has no cron-jobs.json, standalone unit test
  on _restore_cron_skill_links exercising the full report shape.

Validation
- 484/484 targeted tests pass (curator + cron + skill_manager suites).
- E2E: real snapshot_skills, real cron rewrite, real rollback. Before:
  ['pr-review-format', 'pr-review-checklist', 'pr-triage-salvage'].
  After curator: ['hermes-agent-dev']. After rollback: ['pr-review-format',
  'pr-review-checklist', 'pr-triage-salvage']. Non-skill fields (id,
  name, prompt) preserved across the round trip.
2026-05-02 01:29:57 -07:00

431 lines
15 KiB
Python

"""CLI subcommand: `hermes curator <subcommand>`.
Thin shell around agent/curator.py and tools/skill_usage.py. Renders a status
table, triggers a run, pauses/resumes, and pins/unpins skills.
This module intentionally has no side effects at import time — main.py wires
the argparse subparsers on demand.
"""
from __future__ import annotations
import argparse
import sys
from datetime import datetime, timezone
from typing import Optional
def _fmt_ts(ts: Optional[str]) -> str:
if not ts:
return "never"
try:
dt = datetime.fromisoformat(ts)
except (TypeError, ValueError):
return str(ts)
if dt.tzinfo is None:
dt = dt.replace(tzinfo=timezone.utc)
delta = datetime.now(timezone.utc) - dt
secs = int(delta.total_seconds())
if secs < 60:
return f"{secs}s ago"
if secs < 3600:
return f"{secs // 60}m ago"
if secs < 86400:
return f"{secs // 3600}h ago"
return f"{secs // 86400}d ago"
def _cmd_status(args) -> int:
from agent import curator
from tools import skill_usage
state = curator.load_state()
enabled = curator.is_enabled()
paused = state.get("paused", False)
last_run = state.get("last_run_at")
summary = state.get("last_run_summary") or "(none)"
runs = state.get("run_count", 0)
status_line = (
"ENABLED" if enabled and not paused else
"PAUSED" if paused else
"DISABLED"
)
print(f"curator: {status_line}")
print(f" runs: {runs}")
print(f" last run: {_fmt_ts(last_run)}")
print(f" last summary: {summary}")
_report = state.get("last_report_path")
if _report:
print(f" last report: {_report}")
_ih = curator.get_interval_hours()
_interval_label = (
f"{_ih // 24}d" if _ih % 24 == 0 and _ih >= 24
else f"{_ih}h"
)
print(f" interval: every {_interval_label}")
print(f" stale after: {curator.get_stale_after_days()}d unused")
print(f" archive after: {curator.get_archive_after_days()}d unused")
rows = skill_usage.agent_created_report()
if not rows:
print("\nno agent-created skills")
return 0
by_state = {"active": [], "stale": [], "archived": []}
pinned = []
for r in rows:
state_name = r.get("state", "active")
by_state.setdefault(state_name, []).append(r)
if r.get("pinned"):
pinned.append(r["name"])
print(f"\nagent-created skills: {len(rows)} total")
for state_name in ("active", "stale", "archived"):
bucket = by_state.get(state_name, [])
print(f" {state_name:10s} {len(bucket)}")
if pinned:
print(f"\npinned ({len(pinned)}): {', '.join(pinned)}")
# Show top 5 least-recently-active skills. Views and edits are activity too:
# curator should not report a skill as "never used" right after skill_view()
# or skill_manage() touched it.
active = sorted(
by_state.get("active", []),
key=lambda r: r.get("last_activity_at") or r.get("created_at") or "",
)[:5]
if active:
print("\nleast recently active (top 5):")
for r in active:
last = _fmt_ts(r.get("last_activity_at"))
print(
f" {r['name']:40s} "
f"activity={r.get('activity_count', 0):3d} "
f"use={r.get('use_count', 0):3d} "
f"view={r.get('view_count', 0):3d} "
f"patches={r.get('patch_count', 0):3d} "
f"last_activity={last}"
)
# Show top 5 most-active and least-active skills by activity_count
# (use + view + patch). This is a different signal from
# least-recently-active: activity_count reflects frequency,
# last_activity_at reflects recency. A skill touched 30 times a year
# ago is high-frequency but stale; a skill touched once yesterday is
# recent but low-frequency. Both can matter.
active_all = by_state.get("active", [])
if active_all:
most_active = sorted(
active_all,
key=lambda r: (r.get("activity_count") or 0, r.get("last_activity_at") or ""),
reverse=True,
)[:5]
if most_active and (most_active[0].get("activity_count") or 0) > 0:
print("\nmost active (top 5):")
for r in most_active:
last = _fmt_ts(r.get("last_activity_at"))
print(
f" {r['name']:40s} "
f"activity={r.get('activity_count', 0):3d} "
f"use={r.get('use_count', 0):3d} "
f"view={r.get('view_count', 0):3d} "
f"patches={r.get('patch_count', 0):3d} "
f"last_activity={last}"
)
least_active = sorted(
active_all,
key=lambda r: (r.get("activity_count") or 0, r.get("last_activity_at") or ""),
)[:5]
if least_active:
print("\nleast active (top 5):")
for r in least_active:
last = _fmt_ts(r.get("last_activity_at"))
print(
f" {r['name']:40s} "
f"activity={r.get('activity_count', 0):3d} "
f"use={r.get('use_count', 0):3d} "
f"view={r.get('view_count', 0):3d} "
f"patches={r.get('patch_count', 0):3d} "
f"last_activity={last}"
)
return 0
def _cmd_run(args) -> int:
from agent import curator
if not curator.is_enabled():
print("curator: disabled via config; enable with `curator.enabled: true`")
return 1
dry = bool(getattr(args, "dry_run", False))
if dry:
print("curator: running DRY-RUN (report only, no mutations)...")
else:
print("curator: running review pass...")
def _on_summary(msg: str) -> None:
print(msg)
result = curator.run_curator_review(
on_summary=_on_summary,
synchronous=bool(args.synchronous),
dry_run=dry,
)
auto = result.get("auto_transitions", {})
if auto:
if dry:
print(
f"auto (preview): {auto.get('checked', 0)} candidate skill(s) "
"— no transitions applied in dry-run"
)
else:
print(
f"auto: checked={auto.get('checked', 0)} "
f"stale={auto.get('marked_stale', 0)} "
f"archived={auto.get('archived', 0)} "
f"reactivated={auto.get('reactivated', 0)}"
)
if not args.synchronous:
print("llm pass running in background — check `hermes curator status` later")
if dry:
print(
"dry-run: no changes applied. When the report lands, read it with "
"`hermes curator status` and run `hermes curator run` (no flag) to apply."
)
return 0
def _cmd_pause(args) -> int:
from agent import curator
curator.set_paused(True)
print("curator: paused")
return 0
def _cmd_resume(args) -> int:
from agent import curator
curator.set_paused(False)
print("curator: resumed")
return 0
def _cmd_pin(args) -> int:
from tools import skill_usage
if not skill_usage.is_agent_created(args.skill):
print(
f"curator: '{args.skill}' is bundled or hub-installed — cannot pin "
"(only agent-created skills participate in curation)"
)
return 1
skill_usage.set_pinned(args.skill, True)
print(f"curator: pinned '{args.skill}' (will bypass auto-transitions)")
return 0
def _cmd_unpin(args) -> int:
from tools import skill_usage
if not skill_usage.is_agent_created(args.skill):
print(
f"curator: '{args.skill}' is bundled or hub-installed — "
"there's nothing to unpin (curator only tracks agent-created skills)"
)
return 1
skill_usage.set_pinned(args.skill, False)
print(f"curator: unpinned '{args.skill}'")
return 0
def _cmd_restore(args) -> int:
from tools import skill_usage
ok, msg = skill_usage.restore_skill(args.skill)
print(f"curator: {msg}")
return 0 if ok else 1
def _cmd_backup(args) -> int:
"""Take a manual snapshot of the skills tree. Same mechanism as the
automatic pre-run snapshot, just user-initiated."""
from agent import curator_backup
if not curator_backup.is_enabled():
print(
"curator: backups are disabled via config "
"(`curator.backup.enabled: false`); re-enable to snapshot"
)
return 1
reason = getattr(args, "reason", None) or "manual"
snap = curator_backup.snapshot_skills(reason=reason)
if snap is None:
print("curator: snapshot failed — check logs (backup disabled or IO error)")
return 1
print(f"curator: snapshot created at ~/.hermes/skills/.curator_backups/{snap.name}")
return 0
def _cmd_rollback(args) -> int:
"""Restore the skills tree from a snapshot. Defaults to newest.
``--list`` prints available snapshots and exits. ``--id <stamp>`` picks
a specific one. Without ``-y``, prompts for confirmation. A safety
snapshot of the current tree is always taken first, so rollbacks are
themselves undoable.
"""
from agent import curator_backup
if getattr(args, "list", False):
print(curator_backup.summarize_backups())
return 0
backup_id = getattr(args, "backup_id", None)
target_path = curator_backup._resolve_backup(backup_id)
if target_path is None:
rows = curator_backup.list_backups()
if not rows:
print(
"curator: no snapshots exist yet. Take one with "
"`hermes curator backup` or wait for the next curator run."
)
else:
print(
f"curator: no snapshot matching "
f"{'id ' + repr(backup_id) if backup_id else 'your query'}."
)
print("Available:")
print(curator_backup.summarize_backups())
return 1
manifest = curator_backup._read_manifest(target_path)
print(f"Rollback target: {target_path.name}")
if manifest:
print(f" reason: {manifest.get('reason', '?')}")
print(f" created_at: {manifest.get('created_at', '?')}")
print(f" skill files: {manifest.get('skill_files', '?')}")
cron = manifest.get("cron_jobs") or {}
if isinstance(cron, dict):
if cron.get("backed_up"):
print(
f" cron jobs: {cron.get('jobs_count', 0)} "
f"(will be restored for skill-link fields only)"
)
else:
reason = cron.get("reason", "not captured")
print(f" cron jobs: not in snapshot ({reason})")
print(
"\nThis will replace the current ~/.hermes/skills/ tree (a safety "
"snapshot of the current state is taken first so this is undoable). "
"Cron jobs that still exist will have their skills/skill fields "
"restored from the snapshot; all other cron fields are left alone."
)
if not getattr(args, "yes", False):
try:
ans = input("Proceed? [y/N] ").strip().lower()
except (EOFError, KeyboardInterrupt):
print("\ncancelled")
return 1
if ans not in ("y", "yes"):
print("cancelled")
return 1
ok, msg, _ = curator_backup.rollback(backup_id=target_path.name)
if ok:
print(f"curator: {msg}")
return 0
print(f"curator: rollback failed — {msg}")
return 1
# ---------------------------------------------------------------------------
# argparse wiring (called from hermes_cli.main)
# ---------------------------------------------------------------------------
def register_cli(parent: argparse.ArgumentParser) -> None:
"""Attach `curator` subcommands to *parent*.
main.py calls this with the ArgumentParser returned by
``subparsers.add_parser("curator", ...)``.
"""
parent.set_defaults(func=lambda a: (parent.print_help(), 0)[1])
subs = parent.add_subparsers(dest="curator_command")
p_status = subs.add_parser("status", help="Show curator status and skill stats")
p_status.set_defaults(func=_cmd_status)
p_run = subs.add_parser("run", help="Trigger a curator review now")
p_run.add_argument(
"--sync", "--synchronous", dest="synchronous", action="store_true",
help="Wait for the LLM review pass to finish (default: background thread)",
)
p_run.add_argument(
"--dry-run", dest="dry_run", action="store_true",
help="Report only — no state changes, no archives, no consolidation "
"(use this to preview what curator would do)",
)
p_run.set_defaults(func=_cmd_run)
p_pause = subs.add_parser("pause", help="Pause the curator until resumed")
p_pause.set_defaults(func=_cmd_pause)
p_resume = subs.add_parser("resume", help="Resume a paused curator")
p_resume.set_defaults(func=_cmd_resume)
p_pin = subs.add_parser("pin", help="Pin a skill so the curator never auto-transitions it")
p_pin.add_argument("skill", help="Skill name")
p_pin.set_defaults(func=_cmd_pin)
p_unpin = subs.add_parser("unpin", help="Unpin a skill")
p_unpin.add_argument("skill", help="Skill name")
p_unpin.set_defaults(func=_cmd_unpin)
p_restore = subs.add_parser("restore", help="Restore an archived skill")
p_restore.add_argument("skill", help="Skill name")
p_restore.set_defaults(func=_cmd_restore)
p_backup = subs.add_parser(
"backup",
help="Take a manual tar.gz snapshot of ~/.hermes/skills/ "
"(curator also does this automatically before every real run)",
)
p_backup.add_argument(
"--reason", default=None,
help="Free-text label stored in manifest.json (default: 'manual')",
)
p_backup.set_defaults(func=_cmd_backup)
p_rollback = subs.add_parser(
"rollback",
help="Restore ~/.hermes/skills/ from a curator snapshot "
"(defaults to the newest)",
)
p_rollback.add_argument(
"--list", action="store_true",
help="List available snapshots and exit without restoring",
)
p_rollback.add_argument(
"--id", dest="backup_id", default=None,
help="Snapshot id to restore (see `--list`); default: newest",
)
p_rollback.add_argument(
"-y", "--yes", action="store_true",
help="Skip confirmation prompt",
)
p_rollback.set_defaults(func=_cmd_rollback)
def cli_main(argv=None) -> int:
"""Standalone entry (also usable by hermes_cli.main fallthrough)."""
parser = argparse.ArgumentParser(prog="hermes curator")
register_cli(parser)
args = parser.parse_args(argv)
fn = getattr(args, "func", None)
if fn is None:
parser.print_help()
return 0
return int(fn(args) or 0)
if __name__ == "__main__": # pragma: no cover
sys.exit(cli_main())