Install
$ agentstack add skill-pfeff-claude-skills-kb-compile ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
KB Compile
The core behavior of the LLM Knowledge Base workflow (SPEC v2 §2.2). Incrementally compiles un-compiled sources from the vault's raw/ queue into the vault's existing note idioms — type:-routed notes in Notes/, cross-linked into Keywords/ and the MOC/Dataview machinery.
> Status: the deterministic primitives are implemented and unit-tested in kb-core > (source_key, is_writable, concept_slug, TYPE_*). Summarization, concept extraction, > and the writes are agent-orchestrated per operations/compile.md.
Billing posture
/kb-compile is operator-invoked and interactive → subscription billing pool. It must never be wired to cron, claude -p, or the Agent SDK (that moves the workflow to the metered pool; requires operator sign-off). See SPEC v2 §6 and the CLAUDE.md billing guard.
What it writes (SPEC v2 §2.2)
- Source summary →
Notes/YYYY/MM/(type: reference,project: knowledge-base,
sources: []): one summary note per source.
- Concept notes →
Notes/YYYY/MM/(type: zettel): fuller-than-keyword treatments;
created new or updated in place (dedup by concept_slug) on re-compile.
- Keywords/ cross-link: concept notes
[[link]]to term pages; compile may create a new
Keywords/ stub or elaborate an existing one (free-edit, git-reviewable).
- KB MOC (
type: moc) with a Dataview block over the KB notes; backlinks summary ↔
concept ↔ keyword.
Bounded writes: never DevOps Documentation//Confluence/ or notes outside the compile's scope (kb_core.is_writable floor + task-scope). Incremental/idempotent (AC-2.2). Every note carries type: + sources: (INV-2/INV-3). No fence — edit safety is git + review (INV-5).
Invocation
/kb-compile # compile all uncompiled raw/ sources
Execution
- Load the operation:
Read(${CLAUDE_PLUGIN_ROOT}/skills/kb-compile/operations/compile.md) - Execute it.
- Report summaries written, concept notes created/updated,
Keywords/pages touched, and sources compiled.
Integration Points
- kb-core —
${CLAUDE_PLUGIN_ROOT}/skills/kb-core/scripts/kb_core.py(source_key,is_writable,concept_slug,KB_PROJECT,TYPE_*). - obsidian-notes skill / host config — vault path resolution + writes.
- kb-capture — fills the
raw/queue this skill drains. - kb-lint — health-checks the KB notes this skill produces.
See Also
SPEC.mdv2 §1 (ownership model) and §2.2 (compile contract) — workspace.skills/kb-capture,skills/kb-lint.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: pfeff
- Source: pfeff/claude-skills
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.