skillwiki 0.2.1-beta.2 → 0.2.1-beta.4
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/package.json
CHANGED
|
@@ -1,9 +1,8 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "skillwiki",
|
|
3
|
-
"version": "0.2.1-beta.
|
|
4
|
-
"version": "0.2.1-beta.2",
|
|
3
|
+
"version": "0.2.1-beta.4",
|
|
5
4
|
"skills": "./",
|
|
6
|
-
"description": "Project-aware Karpathy-style knowledge base for Claude Code:
|
|
5
|
+
"description": "Project-aware Karpathy-style knowledge base for Claude Code: 15 prompt-only skills (wiki-*, proj-*, using-skillwiki) backed by the deterministic `skillwiki` CLI.",
|
|
7
6
|
"author": {
|
|
8
7
|
"name": "karlorz",
|
|
9
8
|
"url": "https://github.com/karlorz"
|
package/skills/package.json
CHANGED
|
@@ -20,6 +20,7 @@ Invoke a skillwiki skill when the user:
|
|
|
20
20
|
- Wants a health check or lint on their vault
|
|
21
21
|
- Mentions crystallizing a session into a note
|
|
22
22
|
- Talks about project workspaces, ADRs, or distillation
|
|
23
|
+
- Wants to quickly capture an idea, bug, task, or note without interrupting their workflow
|
|
23
24
|
- Wants to archive or clean up old vault pages
|
|
24
25
|
- Needs to detect source drift or re-ingest updated content
|
|
25
26
|
- Has a spec/plan in a non-skillwiki format (CodeStable, RFC, AIDE)
|
|
@@ -27,15 +28,15 @@ Invoke a skillwiki skill when the user:
|
|
|
27
28
|
|
|
28
29
|
## Vault Structure
|
|
29
30
|
|
|
30
|
-
A skillwiki vault has
|
|
31
|
+
A skillwiki vault has three layers. The canonical architecture lives in `SCHEMA.md` at the vault root — read it before creating any new directories.
|
|
31
32
|
|
|
32
|
-
**Layer 1 — Raw (`raw/`):** Immutable source material. Never modify after ingest.
|
|
33
|
+
**Layer 1 — Raw (`raw/`):** Immutable source material. Never modify after ingest. `raw/transcripts/` doubles as the ad-hoc capture point for meeting notes and unprocessed ideas.
|
|
33
34
|
|
|
34
35
|
```
|
|
35
36
|
raw/
|
|
36
37
|
├── articles/ # Web articles, clippings
|
|
37
38
|
├── papers/ # PDFs, arxiv papers
|
|
38
|
-
├── transcripts/ # Meeting notes, interviews
|
|
39
|
+
├── transcripts/ # Meeting notes, interviews, ad-hoc captures
|
|
39
40
|
└── assets/ # Images, diagrams referenced by sources
|
|
40
41
|
```
|
|
41
42
|
|
|
@@ -48,7 +49,19 @@ sha256: # computed by skillwiki hash over body bytes after closing ---
|
|
|
48
49
|
---
|
|
49
50
|
```
|
|
50
51
|
|
|
51
|
-
**Layer 2 —
|
|
52
|
+
**Layer 2 — Typed Knowledge:** `entities/`, `concepts/`, `comparisons/`, `queries/`, `meta/`. Agent-owned pages with `^[raw/...]` citation markers at paragraph-end. Global scope — project association via `provenance_projects:` frontmatter, not directory nesting.
|
|
53
|
+
|
|
54
|
+
**Layer 3 — Project Workspaces (`projects/{slug}/`):** Per-project lifecycle directories with `work/` (spec + plan + retro), `compound/` (distilled lessons/patterns), `architecture/` (ADRs), and `history/` (archived specs/plans).
|
|
55
|
+
|
|
56
|
+
**No `inbox/` directory.** Ad-hoc captures go to `raw/transcripts/` or directly into a project work item via `proj-work`. Do not invent new top-level directories — extend Layer 2 via SCHEMA.md tag taxonomy if needed.
|
|
57
|
+
|
|
58
|
+
### Ad-hoc capture: three entry points
|
|
59
|
+
|
|
60
|
+
| Entry | When | What happens |
|
|
61
|
+
|-------|------|-------------|
|
|
62
|
+
| `/wiki-add-task <text>` | You're in a Claude session | Appends entry to `raw/transcripts/YYYY-MM-DD-ad-hoc-captures.md` |
|
|
63
|
+
| Filesystem drop | You're NOT in a Claude session (Obsidian, editor, sync) | Create/edit any `.md` file in `raw/transcripts/` — dev-loop discovers it on next cycle |
|
|
64
|
+
| Dev-loop discovery | Automatic, next cycle | Scans `raw/transcripts/` for new files since last cycle, surfaces as claimable work |
|
|
52
65
|
|
|
53
66
|
## Skill Map
|
|
54
67
|
|
|
@@ -62,6 +75,7 @@ sha256: # computed by skillwiki hash over body bytes after closing ---
|
|
|
62
75
|
| `wiki-audit` | Verify raw provenance references and source frontmatter integrity |
|
|
63
76
|
| `wiki-archive` | Archive a typed-knowledge page — move to `_archive/`, remove from index |
|
|
64
77
|
| `wiki-reingest` | Detect drift in raw sources (sha256 comparison) and re-ingest updated content |
|
|
78
|
+
| `wiki-add-task` | Quick-capture ideas, bugs, tasks, notes into `raw/transcripts/` without leaving the current workflow |
|
|
65
79
|
| `wiki-adapter-prd` | Map foreign PRD formats (CodeStable, RFC, AIDE, Hermes) into vault pages |
|
|
66
80
|
| `proj-init` | Bootstrap a project workspace (README, requirements, architecture) |
|
|
67
81
|
| `proj-work` | Open or run a work item under a project's work/ directory |
|
|
@@ -0,0 +1,104 @@
|
|
|
1
|
+
---
|
|
2
|
+
name: wiki-add-task
|
|
3
|
+
description: Capture ad-hoc ideas, bugs, tasks, or notes into the vault via /wiki-add-task or filesystem drop.
|
|
4
|
+
---
|
|
5
|
+
|
|
6
|
+
# wiki-add-task
|
|
7
|
+
|
|
8
|
+
Capture ad-hoc ideas, bugs, tasks, and notes into the vault. Three entry points depending on where you are:
|
|
9
|
+
|
|
10
|
+
| Entry | When | What happens |
|
|
11
|
+
|-------|------|-------------|
|
|
12
|
+
| `/wiki-add-task <text>` | You're in a Claude session | Appends entry to `raw/transcripts/YYYY-MM-DD-ad-hoc-captures.md` |
|
|
13
|
+
| Filesystem drop | You're NOT in a Claude session (Obsidian, editor, sync) | Create/edit any file in `raw/transcripts/` — dev-loop discovers it on next cycle |
|
|
14
|
+
| Dev-loop discovery | Automatic, next cycle | Scans `raw/transcripts/` for new files since last cycle, surfaces as claimable work |
|
|
15
|
+
|
|
16
|
+
## When This Skill Activates
|
|
17
|
+
|
|
18
|
+
- User invokes `/wiki-add-task` with a description.
|
|
19
|
+
- User says "add task", "capture this", "note this", "remember this", "log this idea", or similar.
|
|
20
|
+
- User provides a short text description and optionally a type tag.
|
|
21
|
+
|
|
22
|
+
## Output language
|
|
23
|
+
|
|
24
|
+
Run `skillwiki lang` at the start. Entry prose and `--human` summaries use the resolved language. Frontmatter keys, file names, and structural markers stay English.
|
|
25
|
+
|
|
26
|
+
## Steps
|
|
27
|
+
|
|
28
|
+
0. **Resolve vault and language.** Run `skillwiki path` (fail if NO_VAULT_CONFIGURED) and `skillwiki lang`.
|
|
29
|
+
1. **Parse arguments.** Extract from the user's message:
|
|
30
|
+
- `text` — the idea/bug/task/note content (required)
|
|
31
|
+
- `type` — one of: `idea`, `bug`, `task`, `note` (default: `idea`)
|
|
32
|
+
- `project` — optional project slug to cross-reference (e.g., `llm-wiki`)
|
|
33
|
+
2. **Determine target file.** The capture file is `raw/transcripts/YYYY-MM-DD-ad-hoc-captures.md` where YYYY-MM-DD is today's date. If the file exists, append; otherwise create it with standard raw frontmatter.
|
|
34
|
+
3. **Write the entry.** Append to the capture file:
|
|
35
|
+
```markdown
|
|
36
|
+
### HH:MM — [type]
|
|
37
|
+
|
|
38
|
+
[text]
|
|
39
|
+
|
|
40
|
+
<!---meta: {"captured_at": "YYYY-MM-DDTHH:MM:SS", "type": "[type]"}--->
|
|
41
|
+
```
|
|
42
|
+
- Use 24-hour time for HH:MM.
|
|
43
|
+
- Do not overwrite or modify existing entries.
|
|
44
|
+
4. **Cross-reference (optional).** If a `project` slug was provided:
|
|
45
|
+
- Check that `projects/{slug}/` exists in the vault.
|
|
46
|
+
- Append a one-line reference to the project's work log or compound notes:
|
|
47
|
+
`- [YYYY-MM-DD] capture: [text] → raw/transcripts/YYYY-MM-DD-ad-hoc-captures.md`
|
|
48
|
+
- Do NOT create a full work item (that's `proj-work`'s job).
|
|
49
|
+
5. **Update log.md.** Append: `## [YYYY-MM-DD] capture | [type]: [text (first 60 chars)]`
|
|
50
|
+
6. **Confirm to user.** Report what was captured and where. Suggest next steps:
|
|
51
|
+
- If `type: idea` → "Consider ingesting related sources to develop this idea."
|
|
52
|
+
- If `type: bug` → "Use proj-work to create a bug-fix work item."
|
|
53
|
+
- If `type: task` → "Use proj-work to track this task through the dev loop."
|
|
54
|
+
- If `type: note` → "Will be available for future wiki-query searches."
|
|
55
|
+
|
|
56
|
+
## Ad-hoc captures file format
|
|
57
|
+
|
|
58
|
+
The file `raw/transcripts/YYYY-MM-DD-ad-hoc-captures.md` is a standard raw source with frontmatter:
|
|
59
|
+
|
|
60
|
+
```yaml
|
|
61
|
+
---
|
|
62
|
+
source_url:
|
|
63
|
+
ingested: YYYY-MM-DD
|
|
64
|
+
sha256:
|
|
65
|
+
---
|
|
66
|
+
```
|
|
67
|
+
|
|
68
|
+
The `sha256` is computed over the body after the closing `---`. On each append, recompute and update `sha256`. This keeps source-drift detection functional even though the file grows throughout the day.
|
|
69
|
+
|
|
70
|
+
## Stop conditions
|
|
71
|
+
|
|
72
|
+
- `skillwiki path` returns NO_VAULT_CONFIGURED.
|
|
73
|
+
- No `text` provided (prompt user once, then stop).
|
|
74
|
+
|
|
75
|
+
## Forbidden
|
|
76
|
+
|
|
77
|
+
- Creating an `inbox/` directory. All captures go to `raw/transcripts/`.
|
|
78
|
+
- Modifying existing entries in the captures file — only append.
|
|
79
|
+
- Creating a work item — this is capture-only. Use `proj-work` for full work items.
|
|
80
|
+
- Writing to any Layer 2 or Layer 3 location. Captures are Layer 1 (raw).
|
|
81
|
+
|
|
82
|
+
## Filesystem drop (offline capture)
|
|
83
|
+
|
|
84
|
+
When you're not in a Claude session, drop files directly into `raw/transcripts/`:
|
|
85
|
+
|
|
86
|
+
1. Create any `.md` file in `raw/transcripts/` — name it descriptively (e.g., `2026-05-07-idea-xyz.md`)
|
|
87
|
+
2. Add raw frontmatter at the top:
|
|
88
|
+
```yaml
|
|
89
|
+
---
|
|
90
|
+
source_url:
|
|
91
|
+
ingested: YYYY-MM-DD
|
|
92
|
+
sha256:
|
|
93
|
+
---
|
|
94
|
+
```
|
|
95
|
+
3. Write your idea/bug/task/note below the frontmatter
|
|
96
|
+
|
|
97
|
+
No special format required — the dev-loop QUERY step will discover new files on the next cycle and surface them as claimable work. Mark the type with a heading like `## idea`, `## bug`, `## task`, or just write freeform.
|
|
98
|
+
|
|
99
|
+
## Dev-loop discovery
|
|
100
|
+
|
|
101
|
+
When the dev-loop QUERY step runs, it should scan `raw/transcripts/` for files with `ingested:` date newer than the last cycle. New files are surfaced as claimable work items. The agent then decides whether to:
|
|
102
|
+
- Create a work item via `proj-work` (for tasks and bugs)
|
|
103
|
+
- Ingest as a knowledge page via `wiki-ingest` (for ideas with sources)
|
|
104
|
+
- Leave in place (for notes that don't need action yet)
|