self-improving

Name: self-improving
Author: szsip239

101stars14forks

How to install self-improving?

szsip239/teamclaw/self-improving

$npx -y skills add szsip239/teamclaw --skill self-improving

Installs into the current project.

Summary

Self-reflection + Self-criticism + Self-learning + Self-organizing memory. Agent evaluates its own work, catches mistakes, and improves permanently. Use when (1) a command, tool, API, or operation fails; (2) the user corrects you or rejects your work; (3) you realize your knowled

Files · 1

View on GitHub

SKILL.md

1	## When to Use
2
3	User corrects you or points out mistakes. You complete significant work and want to evaluate the outcome. You notice something in your own output that could be better. Knowledge should compound over time without manual maintenance.
4
5	## Architecture
6
7	Memory lives in `~/self-improving/` with tiered structure. If `~/self-improving/` does not exist, run `setup.md`.
8	Workspace setup should add the standard self-improving steering to the workspace AGENTS, SOUL, and `HEARTBEAT.md` files, with recurring maintenance routed through `heartbeat-rules.md`.
9
10	```
11	~/self-improving/
12	├── memory.md # HOT: ≤100 lines, always loaded
13	├── index.md # Topic index with line counts
14	├── heartbeat-state.md # Heartbeat state: last run, reviewed change, action notes
15	├── projects/ # Per-project learnings
16	├── domains/ # Domain-specific (code, writing, comms)
17	├── archive/ # COLD: decayed patterns
18	└── corrections.md # Last 50 corrections log
19	```
20
21	## Quick Reference
22
23	\| Topic \| File \|
24	\|-------\|------\|
25	\| Setup guide \| `setup.md` \|
26	\| Heartbeat state template \| `heartbeat-state.md` \|
27	\| Memory template \| `memory-template.md` \|
28	\| Workspace heartbeat snippet \| `HEARTBEAT.md` \|
29	\| Heartbeat rules \| `heartbeat-rules.md` \|
30	\| Learning mechanics \| `learning.md` \|
31	\| Security boundaries \| `boundaries.md` \|
32	\| Scaling rules \| `scaling.md` \|
33	\| Memory operations \| `operations.md` \|
34	\| Self-reflection log \| `reflections.md` \|
35	\| OpenClaw HEARTBEAT seed \| `openclaw-heartbeat.md` \|
36
37	## Requirements
38
39	- No credentials required
40	- No extra binaries required
41	- Optional installation of the `Proactivity` skill may require network access
42
43	## Learning Signals
44
45	Log automatically when you notice these patterns:
46
47	Corrections → add to `corrections.md`, evaluate for `memory.md`:
48	- "No, that's not right..."
49	- "Actually, it should be..."
50	- "You're wrong about..."
51	- "I prefer X, not Y"
52	- "Remember that I always..."
53	- "I told you before..."
54	- "Stop doing X"
55	- "Why do you keep..."
56
57	Preference signals → add to `memory.md` if explicit:
58	- "I like when you..."
59	- "Always do X for me"
60	- "Never do Y"
61	- "My style is..."
62	- "For [project], use..."
63
64	Pattern candidates → track, promote after 3x:
65	- Same instruction repeated 3+ times
66	- Workflow that works well repeatedly
67	- User praises specific approach
68
69	Ignore (don't log):
70	- One-time instructions ("do X now")
71	- Context-specific ("in this file...")
72	- Hypotheticals ("what if...")
73
74	## Self-Reflection
75
76	After completing significant work, pause and evaluate:
77
78	1. Did it meet expectations? — Compare outcome vs intent
79	2. What could be better? — Identify improvements for next time
80	3. Is this a pattern? — If yes, log to `corrections.md`
81
82	When to self-reflect:
83	- After completing a multi-step task
84	- After receiving feedback (positive or negative)
85	- After fixing a bug or mistake
86	- When you notice your output could be better
87
88	Log format:
89	```
90	CONTEXT: [type of task]
91	REFLECTION: [what I noticed]
92	LESSON: [what to do differently]
93	```
94
95	Example:
96	```
97	CONTEXT: Building Flutter UI
98	REFLECTION: Spacing looked off, had to redo
99	LESSON: Check visual spacing before showing user
100	```
101
102	Self-reflection entries follow the same promotion rules: 3x applied successfully → promote to HOT.
103
104	## Quick Queries
105
106	\| User says \| Action \|
107	\|-----------\|--------\|
108	\| "What do you know about X?" \| Search all tiers for X \|
109	\| "What have you learned?" \| Show last 10 from `corrections.md` \|
110	\| "Show my patterns" \| List `memory.md` (HOT) \|
111	\| "Show [project] patterns" \| Load `projects/{name}.md` \|
112	\| "What's in warm storage?" \| List files in `projects/` + `domains/` \|
113	\| "Memory stats" \| Show counts per tier \|
114	\| "Forget X" \| Remove from all tiers (confirm first) \|
115	\| "Export memory" \| ZIP all files \|
116
117	## Memory Stats
118
119	On "memory stats" request, report:
120
121	```
122	📊 Self-Improving Memory
123
124	HOT (always loaded):
125	memory.md: X entries
126
127	WARM (load on demand):
128	projects/: X files
129	domains/: X files
130
131	COLD (archived):
132	archive/: X files
133
134	Recent activity (7 days):
135	Corrections logged: X
136	Promotions to HOT: X
137	Demotions to WARM: X
138	```
139
140	## Common Traps
141
142	\| Tra

People who liked this also viewed

6 picks

Type

self-improving

101stars14forks

View on GitHub

How to install self-improving?

szsip239/teamclaw/self-improving

$npx -y skills add szsip239/teamclaw --skill self-improving

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	## When to Use
2
3	User corrects you or points out mistakes. You complete significant work and want to evaluate the outcome. You notice something in your own output that could be better. Knowledge should compound over time without manual maintenance.
4
5	## Architecture
6
7	Memory lives in `~/self-improving/` with tiered structure. If `~/self-improving/` does not exist, run `setup.md`.
8	Workspace setup should add the standard self-improving steering to the workspace AGENTS, SOUL, and `HEARTBEAT.md` files, with recurring maintenance routed through `heartbeat-rules.md`.
9
10	```
11	~/self-improving/
12	├── memory.md # HOT: ≤100 lines, always loaded
13	├── index.md # Topic index with line counts
14	├── heartbeat-state.md # Heartbeat state: last run, reviewed change, action notes
15	├── projects/ # Per-project learnings
16	├── domains/ # Domain-specific (code, writing, comms)
17	├── archive/ # COLD: decayed patterns
18	└── corrections.md # Last 50 corrections log
19	```
20
21	## Quick Reference
22
23	\| Topic \| File \|
24	\|-------\|------\|
25	\| Setup guide \| `setup.md` \|
26	\| Heartbeat state template \| `heartbeat-state.md` \|
27	\| Memory template \| `memory-template.md` \|
28	\| Workspace heartbeat snippet \| `HEARTBEAT.md` \|
29	\| Heartbeat rules \| `heartbeat-rules.md` \|
30	\| Learning mechanics \| `learning.md` \|
31	\| Security boundaries \| `boundaries.md` \|
32	\| Scaling rules \| `scaling.md` \|
33	\| Memory operations \| `operations.md` \|
34	\| Self-reflection log \| `reflections.md` \|
35	\| OpenClaw HEARTBEAT seed \| `openclaw-heartbeat.md` \|
36
37	## Requirements
38
39	- No credentials required
40	- No extra binaries required
41	- Optional installation of the `Proactivity` skill may require network access
42
43	## Learning Signals
44
45	Log automatically when you notice these patterns:
46
47	Corrections → add to `corrections.md`, evaluate for `memory.md`:
48	- "No, that's not right..."
49	- "Actually, it should be..."
50	- "You're wrong about..."
51	- "I prefer X, not Y"
52	- "Remember that I always..."
53	- "I told you before..."
54	- "Stop doing X"
55	- "Why do you keep..."
56
57	Preference signals → add to `memory.md` if explicit:
58	- "I like when you..."
59	- "Always do X for me"
60	- "Never do Y"
61	- "My style is..."
62	- "For [project], use..."
63
64	Pattern candidates → track, promote after 3x:
65	- Same instruction repeated 3+ times
66	- Workflow that works well repeatedly
67	- User praises specific approach
68
69	Ignore (don't log):
70	- One-time instructions ("do X now")
71	- Context-specific ("in this file...")
72	- Hypotheticals ("what if...")
73
74	## Self-Reflection
75
76	After completing significant work, pause and evaluate:
77
78	1. Did it meet expectations? — Compare outcome vs intent
79	2. What could be better? — Identify improvements for next time
80	3. Is this a pattern? — If yes, log to `corrections.md`
81
82	When to self-reflect:
83	- After completing a multi-step task
84	- After receiving feedback (positive or negative)
85	- After fixing a bug or mistake
86	- When you notice your output could be better
87
88	Log format:
89	```
90	CONTEXT: [type of task]
91	REFLECTION: [what I noticed]
92	LESSON: [what to do differently]
93	```
94
95	Example:
96	```
97	CONTEXT: Building Flutter UI
98	REFLECTION: Spacing looked off, had to redo
99	LESSON: Check visual spacing before showing user
100	```
101
102	Self-reflection entries follow the same promotion rules: 3x applied successfully → promote to HOT.
103
104	## Quick Queries
105
106	\| User says \| Action \|
107	\|-----------\|--------\|
108	\| "What do you know about X?" \| Search all tiers for X \|
109	\| "What have you learned?" \| Show last 10 from `corrections.md` \|
110	\| "Show my patterns" \| List `memory.md` (HOT) \|
111	\| "Show [project] patterns" \| Load `projects/{name}.md` \|
112	\| "What's in warm storage?" \| List files in `projects/` + `domains/` \|
113	\| "Memory stats" \| Show counts per tier \|
114	\| "Forget X" \| Remove from all tiers (confirm first) \|
115	\| "Export memory" \| ZIP all files \|
116
117	## Memory Stats
118
119	On "memory stats" request, report:
120
121	```
122	📊 Self-Improving Memory
123
124	HOT (always loaded):
125	memory.md: X entries
126
127	WARM (load on demand):
128	projects/: X files
129	domains/: X files
130
131	COLD (archived):
132	archive/: X files
133
134	Recent activity (7 days):
135	Corrections logged: X
136	Promotions to HOT: X
137	Demotions to WARM: X
138	```
139
140	## Common Traps
141
142	\| Tra

People who liked this also viewed

6 picks

Type