ase-help-intent

@${CLAUDE_SKILL_DIR}/../../meta/ase-control.md @${CLAUDE_SKILL_DIR}/../../meta/ase-skill.md @${CLAUDE_SKILL_DIR}/../../meta/ase-dialog.md @${CLAUDE_SKILL_DIR}/../../meta/ase-getopt.md

$ARGUMENTS

The following is the accumulated help of all ASE skills -- the concatenation of every skill's help.md file -- and is the sole catalog you match against:

2. REPEAT the following sub-steps in a LOOP until the user either executes one of the generated commands or cancels the dialog in sub-step 4:

   1. Match Intent:

      Match the current against the and select every skill that adequately fits it -- judging the fit primarily by each skill's ## SCENARIOS ("You want ...") and ## DESCRIPTION sections. Order the selected skills from best-fitting to worst-fitting and keep at most the 8 best ones, so the dispatch dialog of sub-step 4 stays addressable. Set to the number of kept skills. Then, for each kept skill (numbered 1 to in rank order), from that skill's ## SYNOPSIS, ## OPTIONS, and ## ARGUMENTS sections in , generate a concrete command that realizes :

      • Set <name/> to the skill's name (e.g. ase-code-lint).
      • Set <arguments/> to the concrete option flags and positional arguments -- derived from the skill's ## OPTIONS and ## ARGUMENTS -- that best realize (may be empty).
      • Set <command>/ase:<name/> <arguments/></command> (the full command line, with surplus inner spaces collapsed).
      • Set <rationale/> to a very brief, single-sentence justification of why this skill and its options match .

      Finally set yes.

   2. Guard No Match:

   3. Render Commands:

      Output the generated commands, in rank order, with the following , where the [...] marks the repetition of the command/rationale line pair for each kept skill from 1 to :

      C ❯ <command<n/>/> ▷ <rationale/>

      [...]

   4. Dispatch Command:

      In the following, you MUST NOT use your built-in tool! Instead, you MUST just show a custom dialog according to the expanded custom-dialog definition. You MUST closely follow this definition.

      Let the user decide how to proceed by raising a question with the following custom dialog (invoked with --other, so that any free-text instruction is accepted as an intent refinement). Which dialog is raised depends on , so that the C<n/> command options are offered only when commands were actually generated in sub-step 1 and rendered in sub-step 3:

      The [...] line stands for one further answer line C<n/>: Execute <command/> now. per additionally kept skill from 2 to , in rank order, so the dialog offers exactly command options plus CANCEL.

      Check the tool and dispatch accordingly:

      • If is CANCEL: Break out of the loop and stop processing without any further output.

      • If is REFINE, or matches C<n/> while is no: do not execute anything -- output the following and continue the loop at sub-step 4 to obtain a refined intent via the dialog's free-text channel:

      • If matches C<n/> (which implies is yes): Break out of the loop, output the following , and then call the tool Skill(skill: "ase:<name<n/>/>", args: "<arguments<n/>/>") to execute the selected command:

      • If matches OTHER: <text/>: Set (fold the free-text instruction into the intent). Then you MUST continue the loop at sub-step 2.1 to re-match the refined intent.