← ClaudeAtlas

akb-ingestlisted

Ingest whatever you point at into an AKB vault — a local file, a web URL, a GitHub PR/release/commit, a Confluence page, or a Jira issue. Auto-detects the source type and dispatches to a specialized ingest subagent; the router fetches and writes nothing itself. One target per invocation; globs expand to a sequential loop of document ingests.
dnotitia/akb · ★ 56 · AI & Automation · score 71
Install: claude install-skill dnotitia/akb
# AKB Ingest Ingest **one target** into the target AKB vault. Point at a local file, a web URL, a GitHub PR / release / commit, a Confluence page, or a Jira issue — this skill classifies the source type and dispatches to a specialized ingest subagent. It is the single ingest entry point of the `akb-wiki` plugin. **Workflow classification.** LLM-wiki **Ingest** workflow. This is a **router**: it classifies a target and delegates the actual fetch + write to one subagent. Document-style sources pay down **intent debt** (and substrate health) as five-section synthesis pages; decision-record sources (PR / release / commit / issue) pay down **intent debt** as faithful mechanical episodic records. The archetype split (synthesis vs mechanical) is a dispatch decision, invisible to the caller. Scope boundaries: - **One target per invocation.** A glob pattern is the only fan-out: it expands into a sequential loop of `ingest-doc` runs, one per matched file. - **The router fetches nothing and writes nothing.** It only classifies, validates flags, dispatches one subagent, and relays the subagent's report. Each subagent owns exactly one fetch path (file/URL read, `git`, `gh`, or the Atlassian MCP server) — there is no pre-fetched-payload escape hatch at any layer. - **No cross-source wiring at ingest.** Consolidation into concept pages and cross-reference across sources are handled by an external maintenance layer (dream-cycle) over the same vault, not here. ## Arguments ```text /akb-