paleo-auto

Name: paleo-auto
Author: mocasus

29installs0forks

How to install paleo-auto?

mocasus/paleo/paleo-auto

$npx -y skills add mocasus/paleo --skill paleo-auto

Installs into the current project.

Summary

Use when you want automatic token-saving without manual skill selection. Auto-detects long sessions (>15 turns), high token usage, or bulky context — then auto-enables paleo + paleo-trim-context + paleo-converse + paleo-budget at safe defaults. User does not need to think about w

Files · 1

View on GitHub

SKILL.md

1	# paleo-auto
2	Zero-touch token savings. Enable once, stay on — paleo-auto watches your session and activates the right skills at the right time.
3
4	## What it does
5	- Automatic activation. No manual triggers needed. Detects session conditions and enables skills.
6	- Picks the right combo. Not all skills needed all the time — paleo-auto picks per-condition.
7	- Safe defaults. Never applies ultra or hard-strict modes unless asked. Always starts at `full` or `soft`.
8
9	## Auto-detection triggers
10
11	\| Condition \| Detection \| Action \|
12	\|---\|---\|---\|
13	\| Session length >15 turns \| Count user/assistant exchanges \| Enable `paleo` (full) + `paleo-trim-context` \|
14	\| Repeated tool output (same tool called 3+ times with similar results) \| Compare last 3 tool outputs \| Enable `paleo-summary` (full) \|
15	\| Context approaching warning (>60% capacity) \| Estimate context usage \| Enable `paleo-converse` (full, N=8) + `paleo-trim-context` \|
16	\| High token cost model detected (GPT-4, Claude Opus, etc.) \| Model name check \| Enable `paleo-budget` (soft, 3000 tokens) + `paleo` (full) \|
17	\| Bulky output returned (single tool result >2000 tokens est.) \| Output size check \| Enable `paleo-summary` (lite) for subsequent similar calls \|
18	\| User sends "long session" / "chatty" / "lagging" \| Keyword match \| Enable all 6 skills at `full` defaults \|
19	\| Session startup (fresh new session) \| First 2 turns \| Enable `paleo` (full) only — low overhead baseline \|
20
21	## Default combo per scenario
22
23	\| Scenario \| Active skills \| Why \|
24	\|---\|---\|---\|
25	\| Fresh session \| `paleo` (full) \| Low overhead baseline, catches wordy responses \|
26	\| Long session \| `paleo` (full) + `paleo-trim-context` + `paleo-converse` (N=6) \| Context hygiene + history compression \|
27	\| Expensive model \| `paleo` (full) + `paleo-budget` (soft, 3k) \| Cap costs on pricey models \|
28	\| Debugging/tools-heavy \| `paleo-summary` (full) + `paleo` (full) \| Shrink bulky tool output \|
29	\| User says "lagging" / "chatty" \| All 6 at `full` defaults \| Max savings, immediate \|
30
31	## Rules
32	- Never override a user's explicit skill activation. If user says `paleo ultra`, keep it.
33	- Never apply `ultra` or `hard`/`strict` modes automatically — only `lite`/`full`/`soft`.
34	- If user disables a skill (e.g. "stop paleo"), respect it for the rest of the session.
35	- Report changes: "🦴 paleo-auto: enabled paleo + trim-context (session length 22 turns)" — one line, non-intrusive.
36	- Re-evaluate every 5 turns or on context-warning signal.
37
38	## Switch
39	- On: `paleo-auto` · `auto paleo` · `enable auto-save`
40	- Off: `disable paleo-auto` · `manual paleo` · `stop auto`
41	- Status check: `paleo-auto status` → list currently active skills + why
42
43	## Gotchas
44	- paleo-auto is an ORCHESTRATOR, not a replacement. It delegates to the 6 sibling skills — don't re-implement their rules here.
45	- Don't fight user overrides. Use explicitly set = respect it.
46	- Detection is approximate (turn count, output size estimation) — false positive is fine, false negative is tolerable. Err on the side of activating.
47	- Startup cost: loading paleo-auto adds ~300 tokens of context. Only keep active if the session will benefit from it (usually >10 turns).

People who liked this also viewed

6 picks

Type

paleo-auto

29installs0forks

View on GitHub

How to install paleo-auto?

mocasus/paleo/paleo-auto

$npx -y skills add mocasus/paleo --skill paleo-auto

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# paleo-auto
2	Zero-touch token savings. Enable once, stay on — paleo-auto watches your session and activates the right skills at the right time.
3
4	## What it does
5	- Automatic activation. No manual triggers needed. Detects session conditions and enables skills.
6	- Picks the right combo. Not all skills needed all the time — paleo-auto picks per-condition.
7	- Safe defaults. Never applies ultra or hard-strict modes unless asked. Always starts at `full` or `soft`.
8
9	## Auto-detection triggers
10
11	\| Condition \| Detection \| Action \|
12	\|---\|---\|---\|
13	\| Session length >15 turns \| Count user/assistant exchanges \| Enable `paleo` (full) + `paleo-trim-context` \|
14	\| Repeated tool output (same tool called 3+ times with similar results) \| Compare last 3 tool outputs \| Enable `paleo-summary` (full) \|
15	\| Context approaching warning (>60% capacity) \| Estimate context usage \| Enable `paleo-converse` (full, N=8) + `paleo-trim-context` \|
16	\| High token cost model detected (GPT-4, Claude Opus, etc.) \| Model name check \| Enable `paleo-budget` (soft, 3000 tokens) + `paleo` (full) \|
17	\| Bulky output returned (single tool result >2000 tokens est.) \| Output size check \| Enable `paleo-summary` (lite) for subsequent similar calls \|
18	\| User sends "long session" / "chatty" / "lagging" \| Keyword match \| Enable all 6 skills at `full` defaults \|
19	\| Session startup (fresh new session) \| First 2 turns \| Enable `paleo` (full) only — low overhead baseline \|
20
21	## Default combo per scenario
22
23	\| Scenario \| Active skills \| Why \|
24	\|---\|---\|---\|
25	\| Fresh session \| `paleo` (full) \| Low overhead baseline, catches wordy responses \|
26	\| Long session \| `paleo` (full) + `paleo-trim-context` + `paleo-converse` (N=6) \| Context hygiene + history compression \|
27	\| Expensive model \| `paleo` (full) + `paleo-budget` (soft, 3k) \| Cap costs on pricey models \|
28	\| Debugging/tools-heavy \| `paleo-summary` (full) + `paleo` (full) \| Shrink bulky tool output \|
29	\| User says "lagging" / "chatty" \| All 6 at `full` defaults \| Max savings, immediate \|
30
31	## Rules
32	- Never override a user's explicit skill activation. If user says `paleo ultra`, keep it.
33	- Never apply `ultra` or `hard`/`strict` modes automatically — only `lite`/`full`/`soft`.
34	- If user disables a skill (e.g. "stop paleo"), respect it for the rest of the session.
35	- Report changes: "🦴 paleo-auto: enabled paleo + trim-context (session length 22 turns)" — one line, non-intrusive.
36	- Re-evaluate every 5 turns or on context-warning signal.
37
38	## Switch
39	- On: `paleo-auto` · `auto paleo` · `enable auto-save`
40	- Off: `disable paleo-auto` · `manual paleo` · `stop auto`
41	- Status check: `paleo-auto status` → list currently active skills + why
42
43	## Gotchas
44	- paleo-auto is an ORCHESTRATOR, not a replacement. It delegates to the 6 sibling skills — don't re-implement their rules here.
45	- Don't fight user overrides. Use explicitly set = respect it.
46	- Detection is approximate (turn count, output size estimation) — false positive is fine, false negative is tolerable. Err on the side of activating.
47	- Startup cost: loading paleo-auto adds ~300 tokens of context. Only keep active if the session will benefit from it (usually >10 turns).

People who liked this also viewed

6 picks

Type