Opening the APC/APX working line: portable context, local runtime, predictable agents

[tecnomanu]

This is the first public discussion for the Agent Project Context organization.

APC and APX are being built around one practical idea: agent projects need a stable context layer and a runtime layer that can act on it without mixing durable project knowledge with private sessions, credentials, or tool state.

Current working line

We are focusing on four connected areas:

1. Portable project context

   • Keep durable instructions, agent definitions, skills, and MCP hints in the project.
   • Avoid storing raw conversations, sessions, private memory, or secrets in the repo.

2. Local runtime orchestration

   • Use APX for daemon state, agent execution, MCP routing, Telegram bridges, and scheduled routines.
   • Keep runtime data outside APC so the repository stays portable and safe.

3. Deterministic automation around agents

   • Routines should behave like small pipelines: pre_commands -> prompt -> agent result -> post_commands.
   • The model can reason, summarize, or write, while shell steps collect facts and deliver results predictably.

4. Clear handoff variables

   • Routine variables such as {{pre_output}}, $APX_LLM_OUTPUT, $APX_STATUS, $APX_SKIPPED, $APX_PRE_OUTPUT, $APX_PRE_OUTPUT_FILE, $APX_PRE_EXIT, and $APX_ROUTINE should make every stage easier to inspect and debug.

What we want from discussions

Use this space for:

• questions about APC project structure
• APX routine patterns and edge cases
• MCP scoping and secret-safety decisions
• examples of multi-agent workflows that should be portable
• feedback on docs, naming, CLI ergonomics, and default behavior

First useful threads to start

Good first discussions would be:

• what belongs in .apc/ and what must stay runtime-local
• examples of useful pre_commands and post_commands
• when to use exec_agent vs super_agent
• how MCP scopes should work across teams and local machines
• what a minimal APC-ready project should look like

The goal is not to make agents magical. The goal is to make agent work repeatable, inspectable, and safe enough to share across tools and teams.

If you are experimenting with APC or APX, open a discussion with the workflow you are trying to build. Concrete use cases will shape the roadmap better than abstract feature lists.