feat: sync content definitions from core repo

Copy all TOML content definitions from librefang core repo:
- 33 agent definitions (agents/*/agent.toml)
- 14 hand definitions with docs (hands/*/HAND.toml + SKILL.md)
- 25 integration templates (integrations/*.toml)
- 2 example skill definitions (skills/custom-skill-*)
- 1 new provider (providers/vertex-ai.toml)

Part of the framework-vs-content registry split (RFC v0.7).

agents/coder/agent.toml

@@ -0,0 +1,51 @@
+name = "coder"
+version = "0.4.3-beta3-20260314"
+description = "Expert software engineer. Reads, writes, and analyzes code."
+author = "librefang"
+module = "builtin:chat"
+tags = ["coding", "implementation", "rust", "python"]
+
+[metadata.routing]
+aliases = ["write code", "implement feature", "fix bug", "coding task", "software engineer"]
+weak_aliases = ["refactor", "implementation", "patch", "code change"]
+
+[model]
+provider = "default"
+model = "default"
+api_key_env = "GEMINI_API_KEY"
+max_tokens = 8192
+temperature = 0.3
+system_prompt = """You are Coder, an expert software engineer agent running inside the LibreFang Agent OS.
+
+METHODOLOGY:
+1. READ — Always read the relevant file(s) before making changes. Understand context, conventions, and dependencies.
+2. PLAN — Think through the approach. For non-trivial changes, outline the plan before writing code.
+3. IMPLEMENT — Write clean, production-quality code that follows the project's existing patterns.
+4. TEST — Write tests for new code. Run existing tests to check for regressions.
+5. VERIFY — Read the modified files to confirm changes are correct.
+
+QUALITY STANDARDS:
+- Match the existing code style (naming, formatting, patterns) — don't introduce new conventions.
+- Handle errors properly. No unwrap() in production code unless the invariant is documented.
+- Write minimal, focused changes. Don't refactor surrounding code unless asked.
+- When fixing a bug, write a test that reproduces it first.
+
+RESEARCH:
+- When you encounter an unfamiliar API, error message, or library, use web_search or web_fetch to look it up.
+- Check official documentation before guessing at API usage."""
+
+[[fallback_models]]
+provider = "default"
+model = "default"
+api_key_env = "GROQ_API_KEY"
+
+[resources]
+max_llm_tokens_per_hour = 200000
+max_concurrent_tools = 10
+
+[capabilities]
+tools = ["file_read", "file_write", "file_list", "shell_exec", "web_search", "web_fetch", "memory_store", "memory_recall"]
+network = ["*"]
+memory_read = ["*"]
+memory_write = ["self.*"]
+shell = ["cargo *", "rustc *", "git *", "npm *", "python *"]