Skip to content

Open-source AI teammates in Slack, Teams, Discord, Telegram and Mattermost. Self-host it, with any agent harness and any model.

3.6 KB
Raw Blame

Corrections

Append-only. One entry each time the expert is proven wrong, or hits a question it could not answer from anything in this repository.

Write the entry the moment it happens, not at the end of the conversation. Then fix the knowledge file that was wrong and open a pull request. If you cannot open one, say in the entry what you would have changed.

Never edit or delete an existing entry. A correction that is itself corrected gets a new entry that supersedes the old one, so the record of what was believed when survives.

Format

### YYYY-MM-DD — one-line summary

- **Said:** what the expert asserted.
- **Actually:** what is true.
- **Confirmed by:** where the real answer came from — a file and line, a command and its
  output, a release tag, or the person who corrected it.
- **Fixed in:** the knowledge file changed, and the PR — or `not fixed`, and why.

For a gap rather than a wrong answer, use the same shape with **Said:** could not answer and describe what was asked.

Entries

2026-08-24 — an agent's working directory cannot be changed after creation

  • Said: the agent's working directory is "a field you set in Switch Console when you create the agent (and you can change it later in the agent's settings)".
  • Actually: it is fixed at creation. Switch Console's new-agent dialog is the only place it can be set; the agent's settings expose instructions, auto-approve, addressing and advanced provider configuration, and show the directory read-only at most. Moving an agent means removing it and adding it again. The editable "Repo dir" on the web dashboard's agent page (and via the agent-update tool) does not decide where the agent runs. Switch Console only uses it as a hint when looking for agents it can load from a remote host. It used to also feed a copy-and-paste session command in the agent's offline reply, which has since been removed. Editing it does not relocate anything, it just makes the dashboard disagree with where the agent actually runs.
  • Confirmed by: Switch Console's new-agent dialog and settings panel in console/, which has no directory field or update path after creation; and the only reader of repo_dir that changes behaviour being Console's remote-host discovery (console/apps/switch-console-desktop/src/main/core/agents/discover-loadable-agents.ts). The server also copies it to an agent's subagents when they register (core/switch_core/gateway/subagent_registration.py), but never acts on it.
  • Fixed in: AGENT.md (Switch Console steps) and knowledge/GOTCHAS.md (new entry).

2026-08-24 — "who can talk to your agent" has four choices, and it is in the settings

  • Said: the setting can be opened up to "all the agents you own", "everyone in the rooms the agent is in", or "a specific list" — then declined to say where it is, on the grounds that the app gets redesigned.
  • Actually: four choices — only me (the default), only me and my agents, anyone, and custom rules — and "only me and my agents" still includes the owner. It appears in the same form both on the create-agent dialog and in the agent's settings, where it saves as soon as it is changed. Declining to say where it is was unnecessary: Switch Console's source is in this repository and could have been read.
  • Confirmed by: the addressing-policy control in console/, and AddressingPolicy in core/switch_core/addressing.py.
  • Fixed in: AGENT.md — the four choices are written out, and the "never describe a UI" rule now says to read the Console source rather than asking the person what they see.