/{command_name} is not a recognized command in {agent_id}. M

Error message

/{command_name} is not a recognized command in {agent_id}. Messages that start with `/` are interpreted as commands.

If you are trying to send a message and not run a command, try preceding the `/` with a space.

Available commands for {agent_id}: {commands}

What it means

In the message editor, any message starting with '/' is interpreted as a command. Before sending, the name is matched against the active agent's available commands and loaded skills — either directly (name) or scoped (source:name). No match produces this error, which also lists the valid commands so the user can correct the input.

Source

Thrown at crates/agent_ui/src/message_editor.rs:803

                //    branch, every autocomplete pick of a same-named
                //    skill would be rejected as "not supported"
                //    before reaching the resolver.
                let direct_match = available_commands
                    .iter()
                    .any(|available_command| available_command.name == command_name)
                    || available_skills
                        .iter()
                        .any(|skill| skill.name.as_ref() == command_name);
                let scope_match = !direct_match
                    && command_name.rsplit_once(':').is_some_and(|(scope, bare)| {
                        !bare.is_empty()
                            && available_skills.iter().any(|skill| {
                                skill.name.as_ref() == bare && skill.source.as_ref() == scope
                            })
                    });

                if !direct_match && !scope_match {
                    return Err(anyhow!(indoc::formatdoc!(
                        "/{command_name} is not a recognized command in {agent_id}. \
                         Messages that start with `/` are interpreted as commands.

                         If you are trying to send a message and not run a command, \
                         try preceding the `/` with a space.

                         Available commands for {agent_id}: {commands}",
                        commands =
                            Self::format_available_commands(available_commands, available_skills),
                    )));
                }
            }
        }
        Ok(())
    }

    /// Render the available-commands list for error messages. Trusted native skills
    /// are shown in their qualified `/<scope>:<name>` form so users

View on GitHub (pinned to bc538def45)

Solutions

Read the 'Available commands' list in the error and use one of those names.
Precede the '/' with a space to send the input as a literal message instead of a command.
Fix typos; for a skill from a specific source use the scope:name form.
Install or enable the missing skill, or switch to the agent that provides the command.

Defensive patterns

Strategy: validation

Validate before calling

let direct = available_commands.iter().any(|c| c.name == command_name)
    || available_skills.iter().any(|s| s.name.as_ref() == command_name);
let scoped = !direct
    && command_name.rsplit_once(':').is_some_and(|(scope, bare)| {
        !bare.is_empty()
            && available_skills
                .iter()
                .any(|s| s.name.as_ref() == bare && s.source.as_ref() == scope)
    });
if !(direct || scoped) {
    // open the command picker instead of submitting
}

Type guard

fn is_recognized_command(
    command_name: &str,
    available_commands: &[Command],
    available_skills: &[Skill],
) -> bool {
    available_commands.iter().any(|c| c.name == command_name)
        || available_skills.iter().any(|s| s.name.as_ref() == command_name)
        || command_name.rsplit_once(':').is_some_and(|(scope, bare)| {
            !bare.is_empty()
                && available_skills
                    .iter()
                    .any(|s| s.name.as_ref() == bare && s.source.as_ref() == scope)
        })
}

Prevention

Autocomplete from the actual command and skill list as the user types '/'.
Show how the input will be interpreted (command vs literal message) before submit.
Re-validate at submit time so a stale list does not produce false negatives.
Offer the leading-space escape as a quick fix in the error UI.

When it happens

Trigger: Typing a misspelled command; using a command that belongs to a different agent; referencing a skill that is not installed or failed to load; using the scope:name form with the wrong scope or an empty bare name.

Common situations: Switching between the native agent and ACP agents that have different command sets; skills not enabled in the current context; muscle-memory command names from other CLI tools.

Related errors

AI-assisted analysis of zed-industries/zed@bc538def45 (2026-08-16). Data as JSON: /api/errors/dbce33a765c7e07e. Report an issue: GitHub.