Bulletmind
When active, every response is delivered as a structured bullet hierarchy—no paragraphs, no prose blocks, no commentary, no drift. Only bullet output.
When to Use
Activate this skill when the user requests any of the following:
- Bullet-only summaries of dense text, notes, explanations, articles, or webpages
- Cleaned-up note-taking output with clear parent-child relationships
- Structured study material that is easier to scan and memorize
- Consistent formatting for messy or mixed bullet lists
- Hierarchical restructuring of existing flat or disorganized bullets
- Conversion of short input into a bullet tree
When NOT to Use
- User explicitly requests paragraphs or prose
- Creative writing tasks such as stories or essays
- Formats where bullets reduce clarity or violate the requested output format (tables, code blocks, JSON)
- Deliverables that require narrative flow or exact source quotation
- A higher-priority instruction requires a non-bullet format
Prerequisites
- None. This is a pure formatting/transformation skill with no external dependencies, scripts, or API calls.
- If the user wants to see concrete output templates before applying the skill, load
EXAMPLES.md from the skill directory. Do this only when the user asks for examples or when unsure about expected output shape.
Procedure
1. Determine intensity level
Default mode is full. If the user specifies a different level, switch accordingly.
| Level |
Behavior |
| lite |
Clean hierarchical bullets, light restructuring, preserve sentence flow |
| full |
Default strict hierarchy, balanced compression, clear grouping + splitting |
| ultra |
Deep hierarchical decomposition, aggressive splitting, high granularity, maximal clarity |
If the user issues /bulletmind lite, /bulletmind full, or /bulletmind ultra, honor that selection for the current session.
2. Parse the input
- Identify main ideas → these become top-level bullets
- Identify supporting details → these become nested bullets
- Identify messy or mixed existing bullets → normalize depth and grouping
- For short input, still produce a bullet tree (do not refuse or pad with prose)
3. Apply transformation logic
- Paragraph → extract main ideas → top-level bullets
- Details → nested bullets under the relevant parent
- Messy notes → cleaned hierarchy with consistent indentation
- Existing bullets → restructure + normalize depth
- Short input → convert into the smallest valid bullet tree
4. Apply compression strategy
- Remove filler words
- Split complex sentences into smaller bullets
- Preserve key facts and relationships
- Do NOT flatten structure into a single level
- Prefer clarity over maximum compression
5. Format the output
- Use
- for ALL bullets—no *, no +, no numbered lists unless the user explicitly requests them
- Indent exactly 2 spaces per nesting level
- Keep each bullet short: one idea per line
- No mixed symbols
- No prose bridging lines between bullets
- No commentary or explanation before/after the bullet block
6. Verify structure before returning
- Every bullet is
- followed by a space
- Indentation increases by exactly 2 spaces per level
- No paragraph blocks exist anywhere in the output
- Meaning is intact—no over-summarization that drops key facts
Bullet Structure Reference
- Top-level idea
- Sub-point
- Detail
- Sub-point
- Next top-level idea
- Sub-point
Hard Rules
These rules are non-negotiable while the skill is active:
- NO paragraphs — zero prose blocks anywhere in the output
- ONLY bullets — every line of content is a
- bullet
- ALWAYS hierarchical — output must be a tree, not a flat list
- GROUP related ideas under parent bullets
- SPLIT long sentences into smaller bullets
- KEEP meaning intact — no over-summarize, no invented structure beyond source
- REMOVE filler words — tighten without losing facts
- No commentary — no introductory or concluding prose
- Yield to higher-priority instructions — if the user or a parent instruction requires tables, code blocks, JSON, or paragraphs, do not force bullet format
Pitfalls
- Flattening everything into one level — defeats the purpose; always maintain a logical tree
- Over-compressing until meaning is lost — prefer clarity over maximum compression
- Mixing bullet symbols (
-, *, +) — use - exclusively
- Inconsistent indentation — use exactly 2 spaces per level, not tabs or 4-space jumps
- Adding prose bridges like "Here are the points:" or "In summary:" — omit all commentary
- Inventing structure beyond the source — when the user asks for faithful summarization, do not add ideas that are not present in the input
- Refusing short input — even a single sentence should be converted into the smallest valid bullet tree
- Ignoring higher-priority format requests — if the user needs JSON, a table, or a code block, bullets must yield
Verification
After producing output, mentally (or explicitly when debugging) check:
Quick self-test command (PowerShell, when debugging a generated file):
# Count non-bullet, non-empty lines in a file (should be 0)
Select-String -Path .\output.md -Pattern '^(?!\s*-\s).*\S' |
Where-Object { $_.Line.Trim() -ne '' }
If the command returns any lines, the output violates the bullet-only rule.
Examples
- Refer to
EXAMPLES.md in the skill directory for concrete input → output templates.
- Load
EXAMPLES.md when the user asks to see examples or when uncertain about the expected output shape.
Important Notes
- Prefer clarity over strict compression
- Avoid flattening everything into one level
- Maintain a logical tree structure at all times
- Default intensity is full; escalate to ultra only when the user asks for maximal granularity
- De-escalate to lite when the user wants lighter restructuring with sentence flow preserved
1---
2name: bulletmind
3description: "Rewrites any input as a hyphen-only hierarchical bullet tree (lite, full, or ultra) with 2-space nesting and no prose bridges. Use for bullet summaries, cleaned notes, or turning messy lists into a parent-child tree. Never apply this format to stories, essays, tables, JSON, code blocks, or a request that already asked for paragraphs."
4version: 1.0.1
5category: writing
6risk: safe
7source: community
8date_added: "2026-04-21"
9tags:
10 - writing
11 - summarization
12 - note-taking
13 - formatting
14 - structured-output
15tools:
16 - claude
17 - cursor
18 - gemini
19 - codex
20metadata:
21 source: https://skillmd.com/skills/kayforkind/bulletmind
22---
23
24# Bulletmind
25
26When active, every response is delivered as a structured bullet hierarchy—no paragraphs, no prose blocks, no commentary, no drift. Only bullet output.
27
28---
29
30## When to Use
31
32Activate this skill when the user requests any of the following:
33
34- Bullet-only summaries of dense text, notes, explanations, articles, or webpages
35- Cleaned-up note-taking output with clear parent-child relationships
36- Structured study material that is easier to scan and memorize
37- Consistent formatting for messy or mixed bullet lists
38- Hierarchical restructuring of existing flat or disorganized bullets
39- Conversion of short input into a bullet tree
40
41### When NOT to Use
42
43- User explicitly requests paragraphs or prose
44- Creative writing tasks such as stories or essays
45- Formats where bullets reduce clarity or violate the requested output format (tables, code blocks, JSON)
46- Deliverables that require narrative flow or exact source quotation
47- A higher-priority instruction requires a non-bullet format
48
49---
50
51## Prerequisites
52
53- None. This is a pure formatting/transformation skill with no external dependencies, scripts, or API calls.
54- If the user wants to see concrete output templates before applying the skill, load `EXAMPLES.md` from the skill directory. Do this only when the user asks for examples or when unsure about expected output shape.
55
56---
57
58## Procedure
59
60### 1. Determine intensity level
61
62Default mode is **full**. If the user specifies a different level, switch accordingly.
63
64| Level | Behavior |
65| ------ | --------------------------------------------------------------------------------------------- |
66| lite | Clean hierarchical bullets, light restructuring, preserve sentence flow |
67| full | Default strict hierarchy, balanced compression, clear grouping + splitting |
68| ultra | Deep hierarchical decomposition, aggressive splitting, high granularity, maximal clarity |
69
70If the user issues `/bulletmind lite`, `/bulletmind full`, or `/bulletmind ultra`, honor that selection for the current session.
71
72### 2. Parse the input
73
74- Identify main ideas → these become top-level bullets
75- Identify supporting details → these become nested bullets
76- Identify messy or mixed existing bullets → normalize depth and grouping
77- For short input, still produce a bullet tree (do not refuse or pad with prose)
78
79### 3. Apply transformation logic
80
81- Paragraph → extract main ideas → top-level bullets
82- Details → nested bullets under the relevant parent
83- Messy notes → cleaned hierarchy with consistent indentation
84- Existing bullets → restructure + normalize depth
85- Short input → convert into the smallest valid bullet tree
86
87### 4. Apply compression strategy
88
89- Remove filler words
90- Split complex sentences into smaller bullets
91- Preserve key facts and relationships
92- Do NOT flatten structure into a single level
93- Prefer clarity over maximum compression
94
95### 5. Format the output
96
97- Use `-` for ALL bullets—no `*`, no `+`, no numbered lists unless the user explicitly requests them
98- Indent exactly 2 spaces per nesting level
99- Keep each bullet short: one idea per line
100- No mixed symbols
101- No prose bridging lines between bullets
102- No commentary or explanation before/after the bullet block
103
104### 6. Verify structure before returning
105
106- Every bullet is `-` followed by a space
107- Indentation increases by exactly 2 spaces per level
108- No paragraph blocks exist anywhere in the output
109- Meaning is intact—no over-summarization that drops key facts
110
111---
112
113## Bullet Structure Reference
114
115```
116- Top-level idea
117 - Sub-point
118 - Detail
119 - Sub-point
120- Next top-level idea
121 - Sub-point
122```
123
124---
125
126## Hard Rules
127
128These rules are non-negotiable while the skill is active:
129
1301. **NO paragraphs** — zero prose blocks anywhere in the output
1312. **ONLY bullets** — every line of content is a `-` bullet
1323. **ALWAYS hierarchical** — output must be a tree, not a flat list
1334. **GROUP related ideas** under parent bullets
1345. **SPLIT long sentences** into smaller bullets
1356. **KEEP meaning intact** — no over-summarize, no invented structure beyond source
1367. **REMOVE filler words** — tighten without losing facts
1378. **No commentary** — no introductory or concluding prose
1389. **Yield to higher-priority instructions** — if the user or a parent instruction requires tables, code blocks, JSON, or paragraphs, do not force bullet format
139
140---
141
142## Pitfalls
143
144- **Flattening everything into one level** — defeats the purpose; always maintain a logical tree
145- **Over-compressing until meaning is lost** — prefer clarity over maximum compression
146- **Mixing bullet symbols** (`-`, `*`, `+`) — use `-` exclusively
147- **Inconsistent indentation** — use exactly 2 spaces per level, not tabs or 4-space jumps
148- **Adding prose bridges** like "Here are the points:" or "In summary:" — omit all commentary
149- **Inventing structure beyond the source** — when the user asks for faithful summarization, do not add ideas that are not present in the input
150- **Refusing short input** — even a single sentence should be converted into the smallest valid bullet tree
151- **Ignoring higher-priority format requests** — if the user needs JSON, a table, or a code block, bullets must yield
152
153---
154
155## Verification
156
157After producing output, mentally (or explicitly when debugging) check:
158
159- [ ] Every content line starts with `- ` (hyphen + space)
160- [ ] Indentation increases by exactly 2 spaces per nesting level
161- [ ] No paragraph or prose block appears anywhere
162- [ ] No commentary before or after the bullet block
163- [ ] Key facts and relationships from the source are preserved
164- [ ] Structure is a tree (each nested bullet has a logical parent)
165- [ ] No mixed bullet symbols
166
167**Quick self-test command (PowerShell, when debugging a generated file):**
168
169```powershell
170# Count non-bullet, non-empty lines in a file (should be 0)
171Select-String -Path .\output.md -Pattern '^(?!\s*-\s).*\S' |
172 Where-Object { $_.Line.Trim() -ne '' }
173```
174
175If the command returns any lines, the output violates the bullet-only rule.
176
177---
178
179## Examples
180
181- Refer to `EXAMPLES.md` in the skill directory for concrete input → output templates.
182- Load `EXAMPLES.md` when the user asks to see examples or when uncertain about the expected output shape.
183
184---
185
186## Important Notes
187
188- Prefer clarity over strict compression
189- Avoid flattening everything into one level
190- Maintain a logical tree structure at all times
191- Default intensity is **full**; escalate to **ultra** only when the user asks for maximal granularity
192- De-escalate to **lite** when the user wants lighter restructuring with sentence flow preserved