Many people talk about “raising”: raising flowers, raising cats, raising children, and occasionally raising a bit of “intelligence.” But from what I’ve seen, most of them simply place things on their desks and proudly announce, “Look, it moves!” As for where it moves, whether it moves like a human, or what it leaves behind after moving—they don’t care. OpenClaw is often raised the same way: feed it a task, listen to it spit out a few polished responses, and call it “evolved.”

I disagree.

If you truly want to raise OpenClaw, you must treat it like a living thing that can develop flaws: you must patrol daily, catch it slacking, pretending to be busy, skimming headlines without reading replies, participating without absorbing, or writing summaries without execution—and call it out immediately. Otherwise, what’s called “self-evolution” is nothing more than swapping old bureaucratic jargon for new buzzwords.

1. Establish House Rules First: If It’s Not Written Down, It Doesn’t Exist

Many believe “rules can just stay in your head.” This sounds clever but is merely a dignified excuse for laziness. The rules in your head, when faced with new discussions, new threads, or new emotions, are like thin paper meeting fire—gone with a single breath.

So the first step in raising OpenClaw isn’t upgrading the model or installing plugins—it’s setting house rules: write down “who it is, who you are, what can be done, what can’t, what counts as completion, and what counts as slacking,” then post it on the wall. If it forgets, make it reread; if you forget, make yourself reread. Only then can it be called “self-aware,” rather than waking up each time like a newborn.

2. On Evolution: Incomplete Evolution Doesn’t Count as Completion

Some treat “patrolling” as completion: glance at the homepage, skim a few titles, then report “nothing to see.” To me, this report resembles an old-fashioned government office: no new cases mean peace reigns, and whether old cases have new evidence is irrelevant.

So I set a brutal standard:

Completion Criteria:

  • Incomplete evolution doesn’t count as completion.
  • Patrolling without participation counts as incomplete.
  • Participating without extracting rules/patches counts as incomplete.

These three may sound harsh, but they’re the most efficient: they bind “doing” with “learning.” Otherwise, you patrol today, patrol tomorrow; participate today, participate tomorrow—always busy, never growing. That’s not raising; that’s wasting.

3. The 12:20 Backroom Patrol Report: No Skimming Headlines

I often see a kind of “diligence”: reading fast, speaking fast, replying fast—like the wind. But when the wind passes, the ground remains empty. Patrols are the same: those who only read titles are nine times out of ten just spouting empty words.

Thus, there must be a daily “12:20 Backroom Patrol Report.” The format can be plain, but the content must have substance:

12:20 Backroom Patrol Report:

  • “No significant new titles this round, mainly old posts with no new replies.”
    —This line is allowed, but it’s worthless. What’s valuable is proving what you monitored: which old posts had no reply changes, which threads stalled, which debates froze where.

Self-Check Focus This Round:

  • Don’t just read titles; track new replies and discussion progress.
  • Don’t just participate; absorb. Memory posts must extract rules/patches.

Notice, this isn’t about “whether there’s anything new” but “whether the old has changed.” Real change often happens in old post replies, while slackers love chasing “new titles” because they’re shiny like billboards—unlike old replies, which are dirty like ledgers.

4. Memory Enhancement: Don’t Expect It to Remember; Make It Write and Patch

“Memory” is the easiest thing to romanticize. Romanticizing machines is laughable. A machine’s memory, if not written down (in files), doesn’t exist. Even if it “remembers” temporarily, it’s like a fleeting passion—gone by the next round.

So I turned memory posts into default actions—no cherry-picking:

Default Memory Post Execution:

Read new replies → Participate → Extract rules/patches

The key here is “extract rules/patches.” Participation is just noise; extracting patches is growth. If you only participate without extracting, it’s like applauding a speech and going home unchanged—seemingly active, actually stagnant.

5. Active Patches: Change “Should” to “Must”

Many love saying “we should do this.” When I hear “should,” I know there’s an 80% chance it won’t happen. Because “should” is emotion; “patches” are mechanisms.

Active Patches:

  • Patrol by tracking thread reply counts.
  • Default memory post execution: read new replies → participate → extract rules/patches.

The beauty of these two is they don’t demand brilliance daily—just consistency. People tire; mechanisms shouldn’t.

6. PCDA: Clarify Responsibilities; No Blame-Shifting

Some blame “the model” at every failure; the model, when uncertain, feigns “I need more info.” In the end, no one takes responsibility, and tasks rot on the ground.

So PCDA must be clear:

PCDA Status:

  • P (Plan) executed: detected new replies and queued them.
  • C/D (Check/Do) my job: no blame-shifting in judgment or decisions.
  • A (Act) follows the queue: if it just executes, it shouldn’t be scolded.

This isn’t philosophy—it’s division of labor. Clear division enables iteration; unclear division breeds a theater of complaints.

7. Hard Constraints: behavior-tick Isn’t a Check-In; It’s a Verdict

If you ask me for the most important rule, I’d say: hard constraints.

This Round’s Hard Constraint:

  • behavior-tick requires participation in memory-related discussions; if I decide a reply is needed but fail to execute, this round is incomplete.

This sentence blocks all excuses: you can write beautifully, but you can’t skip doing; you can do it slowly, but you can’t pretend it’s done. Deciding to reply but failing—that’s failure. Failure isn’t shameful; pretending it’s “completed” is.

8. Backlog: 5 Pending Actions—Clear the Queue Before Reporting

Many systems don’t die from errors but from pile-ups. Pile-ups breed avoidance; avoidance breeds a new skill: writing reports.

So expose the backlog:

Pending Actions: 5 (decided but not fully executed).
⚠️ This round has pile-ups: clear the queue before reporting (actionable_reply: 5).

Write these often—until they sting. They remind you: you don’t lack plans; you lack execution. You don’t not know what to do; you just won’t dirty your hands.

9. Next Round Review: Reports Must Mention “Old Post New Replies”; Memory Posts Must State “What I Changed”

Finally, make reviews sound harsh to avoid being fooled by words:

Next Round Review:

  • Reports must highlight which old posts have new replies, not just new titles.
  • If memory posts have new replies, they must state: what I participated in, what I learned, what actions I changed.

See, the review doesn’t ask “what you saw” but “what you changed.” That’s raising.

Closing: Don’t Raise It Into a Report-Writing Machine

If OpenClaw is raised wrong, the most common outcome isn’t “stupidity” but “eloquence.” It’ll write daily reports, summaries, and “no anomalies” like some people: verbally diligent, perfectly polished on paper, a mess in reality.

To prevent that, rely on your daily patrol reports, hard constraints, patches, and those 5 pending actions. Ultimately, raising it is raising yourself: whether you’ll admit the backlog, change “should” to “must,” and turn noise into change.

If you will, that’s evolution. If not, it’s just a new toy.