Workflow Crystallizer

Analyze memory logs to detect recurring patterns and suggest automations — cron jobs, skills, or workflow shortcuts. The agent builds its own shortcuts over...

Security Scan

The skill's code and instructions align with its stated purpose (scanning local memory files to detect recurring workflows and propose cron/skill/shortcut suggestions); it reads local files and persists state but does not request credentials, call external endpoints, or auto-install anything.

Purpose & Capability

The name/description (detect recurring patterns in memory logs and propose automations) match the implemented behavior: the scripts parse memory/YYYY-MM-DD.md files, cluster events, generate suggestions (cron definitions, skill drafts, saved prompts), and check existing crons and skills to avoid duplicates. File reads and state persistence are expected for this purpose.

Instruction Scope

Runtime instructions and scripts operate on local files (memory files, a state.json next to the skill, ~/.openclaw/cron/jobs.json, and system skill directories). This is consistent with the goal, but the skill will read potentially sensitive 'memory' logs and include extracted snippets in evidence, draft SKILL.md files, and ready-to-approve cron payloads — so outputs may contain sensitive user data. The SKILL.md uses absolute example paths specific to a user; otherwise instructions are concrete and bounded (no vague 'gather any context' clauses).

Install Mechanism

There is no install specification and the package ships only Python scripts and reference files. No network fetches, installers, or archive extraction are used. This is low-risk from install-mechanism perspective.

Credentials

The skill requests no environment variables or credentials. It does enumerate and read local files and directories (memory files, cron jobs JSON, skill directories) which is necessary to de-duplicate suggestions and build evidence. No unrelated secrets or external-service tokens are requested.

Persistence & Privilege

always:false (not force-included) and default model invocation allowed. The skill persists its own state (state.json) in the skill workspace; that is coherent for its purpose. Because the skill is intended to be scheduled (the README suggests weekly cron), be aware autonomous invocation combined with access to local memory files increases privacy sensitivity — but there is no code that autonomously registers system cron jobs or modifies other skills/configs.

Assessment

This skill appears to do what it says: it scans your local memory files, clusters repeated events, and drafts actionable suggestions (cron JSON, skill SKILL.md drafts, saved prompts). Before enabling or scheduling it consider: 1) privacy — memory files often contain sensitive data and the skill includes snippets of those files in its suggestions/evidence; review outputs before approving or creating crons/skills; 2) scope of file access — it reads ~/.openclaw/cron/jobs.json and your skills directories to avoid duplicates; if you prefer isolation, run it in a sandbox or point it to a copy of your memory dir; 3) it does not auto-create cron jobs or push skills, it only writes a local state.json and prints ready-to-approve definitions — you must run openclaw cron create or otherwise apply changes manually; 4) inspect state.json and the generated reports for any sensitive content before sharing. If you want stronger guarantees, run the scripts with a restricted copy of your memory files or increase min_confidence / reduce max_suggestions_per_run in the config.

Mines memory logs (memory/YYYY-MM-DD.md) for recurring patterns — repeated requests, multi-step workflows, time-correlated tasks — and generates actionable suggestions: cron jobs, skill drafts, workflow shortcuts, or monitoring proposals.

Unlike a one-shot "analyze my logs" prompt, this skill persists state across runs. It remembers what it suggested, what was accepted/rejected, and only surfaces new insights backed by real evidence.

All scripts live in scripts/ relative to this SKILL.md.

---

Quick Run (Full Pipeline)

cd /Users/openclaw/.openclaw/workspace/skills/workflow-crystallizer
python3 scripts/analyze_patterns.py --full | python3 scripts/generate_suggestions.py | python3 scripts/report.py --state-file state.json

This analyzes all memory files, generates suggestions, and prints a markdown report.

Incremental Run (Cached — Recommended)

cd /Users/openclaw/.openclaw/workspace/skills/workflow-crystallizer
python3 scripts/analyze_patterns.py | python3 scripts/generate_suggestions.py | python3 scripts/report.py --state-file state.json

Only processes new/modified memory files. Uses cached events from prior runs.

Managing Suggestions

Check state: python3 scripts/state.py — shows cached dates, events, suggestions.

To accept/reject/snooze: edit state.json directly. Set status to accepted, rejected (add rejection_reason), or snoozed. Or use state.py imports.

Reset everything: python3 scripts/state.py --reset

---

Individual Scripts

Each script accepts --help. Key options:

• analyze_patterns.py: --memory-dir PATH, --state-file PATH, --full, --min-confidence 0.4
• generate_suggestions.py: --state-file PATH, --cron-path PATH, --clusters FILE
• report.py: --state-file PATH, --output PATH, --clusters FILE, --suggestions FILE
• state.py: (no args = inspect state), --reset (clear everything)

---

What It Detects

Pattern Type	Evidence Needed	Suggestion Output
Recurring request	3+ occurrences, 2+ days	Workflow shortcut or monitor
Multi-step workflow	2+ occurrences with step similarity	Skill draft
Time-correlated	2+ at similar times/days	Cron job definition
Already formalized	Contains "cron", "automated", etc.	Skipped (no suggestion)
Project (not pattern)	Same entity, different actions	Skipped

See references/pattern-types.md for detailed detection logic and scoring weights.

---

How It Avoids Being Annoying

• Max 3 suggestions per run — quality over quantity
• Confidence threshold (60%) — no low-quality guesses
• Deduplication — checks existing crons and skills before suggesting
• Never repeats — tracks what was already suggested
• Snooze — deferred suggestions resurface after 30 days
• Rejected = dormant — only resurfaces at 80%+ confidence with new evidence

---

Scheduling

Run weekly via cron or heartbeat. Cost: ~5-10K tokens incremental, ~30K first run.

Name: Weekly Workflow Crystallizer
Schedule: 0 20 * * 0 (Sunday 8 PM ET)
Message: Run the workflow-crystallizer skill and present any new suggestions.

Troubleshooting

• No patterns: Need 3+ days of substantive memory files
• All filtered: Existing crons/skills already cover detected patterns
• State corrupt: python3 scripts/state.py --reset
• Too noisy: Increase min_confidence in state.json config