Skip to content

PLATFORM_SPEC.md — AstrBot

New to gald3r on this platform? Start at Use gald3r with AstrBot for a short overview before the full breakdown below.

Authoring status: create (STUB — generated to end the ghost-platform invisibility; capability cells below were subsequently reconciled with the curated the platform capability matrix / platform_matrix_data.json values). This platform ships an overlay (platforms/astrbot/) and is registered in PLATFORM_REGISTRY.yaml. Hooks/Rules/Skills/Commands/MCP below reflect a curated, human-assessed capability set, not a live docs crawl. Run @g-platform-scan-docs astrbot (crawl https://astrbot.app) then @g-platform-check astrbot to upgrade this spec from curated to verified.

AstrBot is registered in the gald3r platform roster. Per the curated capability matrix (platform_matrix_data.json, reconciled here), the five matrix-tracked primitives — Hooks, Rules, Skills, Commands, and MCP — are all assessed ✅ supported, with an MCP (L2) engine-integration tier and no established rules-file extension (rules_ext: "—"). This is a human-assessed curation, not the result of a live docs crawl or install test: the three primitives the matrix does not track (folder hierarchy, the AI instruction file, and native agent/subagent support) remain ❓ unknown until a live scan runs. This stub exists so the platform is visible to the monitor tooling and the roster-parity gate instead of being silently dropped.

1. Folder Hierarchy

Status: ❓ untested — no verified info. Not a capability the curated matrix tracks either. The platform's config-folder layout has not been crawled/verified. Do not assume a .astrbot/ tree exists until @g-platform-scan-docs astrbot confirms it.

2. AI Instruction File

Status: ❓ untested. Not a capability the curated matrix tracks. Whether AstrBot auto-reads AGENTS.md / CLAUDE.md or a native instruction file is unverified.

3. Agents Support — ❓ untested

Not a matrix-tracked column. No verified info. Pending a live docs crawl.

4. Skills Support — ✅ curated, pending crawl verification

platform_matrix_data.json / the platform capability matrix assess Skills as ✅ supported (human-assessed). No live docs crawl has confirmed the underlying mechanism (SKILL.md discovery path, frontmatter shape, etc.) — treat the ✅ as a curated capability signal, not an implementation-detail source.

5. Commands / Workflows — ✅ curated, pending crawl verification

Curated ✅ supported. Slash-command syntax and workflow file format are unverified — pending @g-platform-scan-docs astrbot.

6. Hooks System — ✅ curated, pending crawl verification

Curated ✅ supported. Native hook config file, event list, and payload format are unverified — pending a live docs crawl.

7. Rules / Memory — ✅ curated, pending crawl verification

Curated ✅ supported. platform_matrix_data.json records rules_ext: "—" (no established rules-file extension curated yet) — do not assume a specific file format until verified.

8. MCP Support — ✅ curated, pending crawl verification

Curated ✅ supported, with a curated engine-integration tier of MCP (L2). Connection config and transport details are unverified — pending a live docs crawl.

Parity vs. Cursor Reference

At the curated level, the five matrix-tracked primitives (Hooks/Rules/Skills/Commands/MCP) match the Cursor reference's ✅ row, i.e. curated full parity on those five columns. This is not a verified parity claim: the three primitives outside the matrix (folder hierarchy, instruction file, agents) are still ❓, and none of the five curated cells has been confirmed by an install test. Re-derive a verified parity comparison against g-skl-platform-cursor/PLATFORM_SPEC.md after @g-platform-scan-docs astrbot runs.

Hook System

  • Type: curated ✅ supported; specific mechanism ❓ untested
  • Config file: unknown
  • Events available: unknown
  • Event payload format: unknown
  • gald3r hook files: portability of g-hk-*.py is unverified

Atypical Handling

  • None recorded — no live scan has run. Capture atypical behaviors after the first verified scan.

gald3r Integration Notes

  • Capability cells (Hooks/Rules/Skills/Commands/MCP) are now curated to match the platform capability matrix. Do not assume any .claude/ reuse, hook, rule, skill, command, or MCP wiring mechanism works on AstrBot until a live docs crawl / install test verifies it — curated only means "assessed as supported," not "implementation confirmed."
  • Run @g-platform-scan-docs astrbot (crawl https://astrbot.app) then @g-platform-check astrbot.

Capability Summary

Hooks Rules Skills Commands MCP Docs Fresh
✅ ✅ ✅ ✅ ✅ ❓

Legend: ✅ verified working · ⚠️ partial / Cursor-generic · ❌ not supported · ❓ untested.

Provenance : the five capability cells above are curated (human-assessed, sourced from platform_matrix_data.json / the platform capability matrix) — they are not live-crawl-verified. Docs Fresh stays ❓ because last_doc_scan: never.

Overall platform status: ❓ (crawl-untested) — capability cells are curated ✅ across the board (reconciliation) but not yet confirmed by a live docs crawl or install test. This is the honest state: curated, not verified.

Verification Evidence

Capability How verified
Hooks / Rules / Skills / Commands / MCP Curated, not crawled. Sourced from platform_matrix_data.json (human-assessed) / the platform capability matrix; reconciled into this spec.
Folder hierarchy / AI instruction file / Agents Not yet verified — no curated or crawled data exists for these three primitives.
(all) Run @g-platform-scan-docs astrbot to replace curated cells with live-crawl-verified ones.

Reference page — canonical source: g-skl-platform-astrbot/PLATFORM_SPEC.md in the engine repo, generated by scripts/generate_platform_support_docs.py. Pages marked curated are hand-assessed and not yet live-verified.