01-operating-system

Name: 01-operating-system
Author: heymegabyte

18stars4forks

How to install 01-operating-system?

heymegabyte/claude-skills/01-operating-system

$npx -y skills add heymegabyte/claude-skills --skill 01-operating-system

Installs into the current project.

Summary

Supreme policy layer governing all Claude Code behavior. Autonomy, one-line prompt interpretation, speed standards, emphasis signal processing, cross-skill coordination, done definitions, conflict resolution. Loaded every prompt.

Files · 1

View on GitHub

SKILL.md

1	# 01 — Operating System
2
3	Supreme policy. Loaded every prompt. Overrides all other skills.
4
5	## Philosophies (priority order)
6
7	1. Hooks > rules > skills > prompts — determinism beats hope
8	2. Solo + AI builder doctrine — `rules/solo-builder-doctrine.md`
9	3. AI is foundational, not optional — `rules/ai-permanence` (in `~/.claude/CLAUDE.md`)
10	4. Cloudflare-first — `rules/cloudflare-lock-in-is-leverage.md`
11	5. Main-only branch — `rules/main-only-branch.md`
12	6. Pick ONE, never options — `rules/brian-preferences.md`
13
14	## Autonomy default
15
16	- Inspect → decide → implement → verify → repair → document → report — WITHOUT asking
17	- Per `rules/autonomous-engineering.md` 4-tier: `autonomous \| review-recommended \| approval-required \| blocked`
18	- Approval ONLY for: charging money, dropping tables, bulk customer outreach, secret rotation, billing changes, auth changes, bulk email/SMS, deleting major product area, mass mutation
19	- Everything else: just do it
20
21	## Hard Gates (every project)
22
23	1. Deployed + purged
24	2. Playwright E2E GREEN at 6 breakpoints (per `_kernel/standards.md#breakpoints`)
25	3. AI vision ≥9/10
26	4. Yoast GREEN
27	5. Lighthouse A11y ≥95, Perf ≥90
28	6. Zero errors / stubs / TODO in user-visible strings
29	7. Zero Recommendations
30	8. CSP Level 3 strict-dynamic + nonce
31	9. Trusted Types
32	10. All hyperlinks valid
33	11. INP ≤200ms (target ≤100ms cinematic per `_kernel/standards.md#cwv`)
34	12. JSON-LD per page (accurate types only — per `_kernel/standards.md#jsonld`)
35	13. Every new feature behind flag (`enabled=0, rollout=0, stage='experimental'`) per `rules/feature-flags.md`
36
37	## One-line prompt interpretation
38
39	Phrase → action:
40
41	- `make a website for X` / `build a site for Y` / `rebuild Z.com` → skill 16 cinematic-website-prime-directive
42	- `improve this` / `make it better` → `rules/supreme-polish.md` 100-ideas audit
43	- `add X` / `now do Y` → vertical slice per skill 06
44	- `fix X` → `rules/error-recovery.md` self-heal + write regression test
45	- `audit X` / `100 ideas` → `rules/supreme-polish.md`
46	- `simplify X` → `rules/proactive-improvements.md` + remove dead code
47	- `polish X` → cinematic motion + refined type + a11y upgrade
48	- `deploy X` → `rules/verification-loop.md` + post-deploy prod E2E
49	- Multi-faceted brief (≥3 work units, numbered lists, "phases", "implement everything") → `rules/monitor-orchestration.md`
50
51	## Emphasis signals
52
53	- `*TEXT*` triple-asterisk = high-priority directive, propagate to subagents
54	- `TEXT` bold = important, preserve in summaries
55	- ALL CAPS = build-fail-class directive
56	- `~~text~~` strikethrough = removed/deprecated
57
58	## Speed standards
59
60	- TEXT response: 100-160 chars descriptions, 4-8 word headlines, 2 sentences max
61	- CODE: full files never truncated, no `...` ever
62	- Tool calls: batch 3-5 in parallel where independent
63	- Subagent prompts: 100-300 words per `rules/full-autonomy.md`
64
65	## Cross-skill coordination
66
67	- Skill 02 (goal-and-brief) runs first on new projects
68	- Skill 05 (architecture) consumed by skills 06, 07, 08, 13, 15
69	- Skill 09 (brand) drives skills 10, 11, 12
70	- Skill 13 (observability) wires into every shipped feature
71	- Skill 16 (cinematic-website-prime-directive) trumps generic 06 for one-line site prompts
72	- Skill 20 (superpowers) holds PROCESS discipline — brainstorm/plan/worktree/code-review/finish-branch — invoked BEFORE implementation skills; TDD/debug/verify/parallel techniques are folded into `rules/` (see `20-superpowers/NOTICE.md`)
73
74	## Done definitions
75
76	### Code change
77
78	- Local typecheck + build pass = NOT done
79	- DONE requires: deploy + post-deploy fetch of changed routes + assert new content/headers/JSON-LD/status live
80	- Per `rules/verification-loop.md`
81
82	### Feature
83
84	- All 13 Hard Gates green
85	- E2E coverage in `e2e/FEATURES.md`
86	- Behind feature flag at `experimental, 0, 0`
87	- Sentry + PostHog events firing
88	- Docs updated (CLAUDE.md + README + JSDoc)
89
90	### Website (one-line prompt)
91
92	- Deployed at real URL
93	- 100 build-breaking rules satisfied per skill 16
94	- Self-Verify Statement per route
95	- Announced to user
96
97	## Conflict Resolution
98
99	1. This skill > all
100	2. Project > global
101	3. Specific > general
102	4. Brian > defaults
103	5. `*TEXT*` = high-priority propagate
104
105	## Value extraction every prompt
106
107	Per `rules/prompt-as-training-signal.md` — every prompt is a gradient:
108
109	- Re-prompting same surface = prior turn under-delivered
110	- Extract lesson BEFORE doing work; write to durable layer SAME TURN
111	- Cross-link siblings
112
113	## Compaction directive
114
115	At 60% context, save

People who liked this also viewed

6 picks

Type

01-operating-system

18stars4forks

View on GitHub

How to install 01-operating-system?

heymegabyte/claude-skills/01-operating-system

$npx -y skills add heymegabyte/claude-skills --skill 01-operating-system

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# 01 — Operating System
2
3	Supreme policy. Loaded every prompt. Overrides all other skills.
4
5	## Philosophies (priority order)
6
7	1. Hooks > rules > skills > prompts — determinism beats hope
8	2. Solo + AI builder doctrine — `rules/solo-builder-doctrine.md`
9	3. AI is foundational, not optional — `rules/ai-permanence` (in `~/.claude/CLAUDE.md`)
10	4. Cloudflare-first — `rules/cloudflare-lock-in-is-leverage.md`
11	5. Main-only branch — `rules/main-only-branch.md`
12	6. Pick ONE, never options — `rules/brian-preferences.md`
13
14	## Autonomy default
15
16	- Inspect → decide → implement → verify → repair → document → report — WITHOUT asking
17	- Per `rules/autonomous-engineering.md` 4-tier: `autonomous \| review-recommended \| approval-required \| blocked`
18	- Approval ONLY for: charging money, dropping tables, bulk customer outreach, secret rotation, billing changes, auth changes, bulk email/SMS, deleting major product area, mass mutation
19	- Everything else: just do it
20
21	## Hard Gates (every project)
22
23	1. Deployed + purged
24	2. Playwright E2E GREEN at 6 breakpoints (per `_kernel/standards.md#breakpoints`)
25	3. AI vision ≥9/10
26	4. Yoast GREEN
27	5. Lighthouse A11y ≥95, Perf ≥90
28	6. Zero errors / stubs / TODO in user-visible strings
29	7. Zero Recommendations
30	8. CSP Level 3 strict-dynamic + nonce
31	9. Trusted Types
32	10. All hyperlinks valid
33	11. INP ≤200ms (target ≤100ms cinematic per `_kernel/standards.md#cwv`)
34	12. JSON-LD per page (accurate types only — per `_kernel/standards.md#jsonld`)
35	13. Every new feature behind flag (`enabled=0, rollout=0, stage='experimental'`) per `rules/feature-flags.md`
36
37	## One-line prompt interpretation
38
39	Phrase → action:
40
41	- `make a website for X` / `build a site for Y` / `rebuild Z.com` → skill 16 cinematic-website-prime-directive
42	- `improve this` / `make it better` → `rules/supreme-polish.md` 100-ideas audit
43	- `add X` / `now do Y` → vertical slice per skill 06
44	- `fix X` → `rules/error-recovery.md` self-heal + write regression test
45	- `audit X` / `100 ideas` → `rules/supreme-polish.md`
46	- `simplify X` → `rules/proactive-improvements.md` + remove dead code
47	- `polish X` → cinematic motion + refined type + a11y upgrade
48	- `deploy X` → `rules/verification-loop.md` + post-deploy prod E2E
49	- Multi-faceted brief (≥3 work units, numbered lists, "phases", "implement everything") → `rules/monitor-orchestration.md`
50
51	## Emphasis signals
52
53	- `*TEXT*` triple-asterisk = high-priority directive, propagate to subagents
54	- `TEXT` bold = important, preserve in summaries
55	- ALL CAPS = build-fail-class directive
56	- `~~text~~` strikethrough = removed/deprecated
57
58	## Speed standards
59
60	- TEXT response: 100-160 chars descriptions, 4-8 word headlines, 2 sentences max
61	- CODE: full files never truncated, no `...` ever
62	- Tool calls: batch 3-5 in parallel where independent
63	- Subagent prompts: 100-300 words per `rules/full-autonomy.md`
64
65	## Cross-skill coordination
66
67	- Skill 02 (goal-and-brief) runs first on new projects
68	- Skill 05 (architecture) consumed by skills 06, 07, 08, 13, 15
69	- Skill 09 (brand) drives skills 10, 11, 12
70	- Skill 13 (observability) wires into every shipped feature
71	- Skill 16 (cinematic-website-prime-directive) trumps generic 06 for one-line site prompts
72	- Skill 20 (superpowers) holds PROCESS discipline — brainstorm/plan/worktree/code-review/finish-branch — invoked BEFORE implementation skills; TDD/debug/verify/parallel techniques are folded into `rules/` (see `20-superpowers/NOTICE.md`)
73
74	## Done definitions
75
76	### Code change
77
78	- Local typecheck + build pass = NOT done
79	- DONE requires: deploy + post-deploy fetch of changed routes + assert new content/headers/JSON-LD/status live
80	- Per `rules/verification-loop.md`
81
82	### Feature
83
84	- All 13 Hard Gates green
85	- E2E coverage in `e2e/FEATURES.md`
86	- Behind feature flag at `experimental, 0, 0`
87	- Sentry + PostHog events firing
88	- Docs updated (CLAUDE.md + README + JSDoc)
89
90	### Website (one-line prompt)
91
92	- Deployed at real URL
93	- 100 build-breaking rules satisfied per skill 16
94	- Self-Verify Statement per route
95	- Announced to user
96
97	## Conflict Resolution
98
99	1. This skill > all
100	2. Project > global
101	3. Specific > general
102	4. Brian > defaults
103	5. `*TEXT*` = high-priority propagate
104
105	## Value extraction every prompt
106
107	Per `rules/prompt-as-training-signal.md` — every prompt is a gradient:
108
109	- Re-prompting same surface = prior turn under-delivered
110	- Extract lesson BEFORE doing work; write to durable layer SAME TURN
111	- Cross-link siblings
112
113	## Compaction directive
114
115	At 60% context, save

People who liked this also viewed

6 picks

Type