commandslisted

# Scope Discoverability index for the seventeen commands in the project command vocabulary. This skill exists so an agent can find out which commands are active without loading all of `working-preferences`. HELP and MAN resolve here. Every other command executes under the definitions `working-preferences` holds. # Command Index - AUDIT [scope]: targeted diagnostic pass. Report errors, gaps, and unverified claims. Do not fix; await instruction. - COMPLETE: finalize now with available context, no further checkpointing. - EXPAND: add depth to the last response without restating it. - EXPORT: recap of session decisions and open threads, pipe-delimited numbered lines (EVENT_00N | CATEGORY | Detail). Append-only across a session. When marking an old entry CLOSED, preserve its original text verbatim rather than rewriting it. - FLAG [content]: treat as load-bearing; surface conflicts before overriding later. - FREEZE [content]: treat as fixed; don't revise without flagging conflict first. - HELP: list active commands. - MAN [command]: full definition of one command. - OVERRIDE: drop an active pushback thread on one point. - PROCEED: current stage is good, advance without recap. - RESET: restate current task, active preferences, and current position. - RETRACT [content]: treat specified content as unsaid going forward. - RETRY: redo the last response in compliance, no acknowledgment of the miss. - REVIEW: re-examine a suspected error in a prior response, explain and revise. - SCOPE: r