callable-sub-summary

name	callable-sub-summary
description	give me a marketing summary. what features can I expect from these callable subs?.
argument-hint	[optional: path-glob for process files]
user-invocable	true

Callable Sub Summary

Generate a concise, marketing-oriented summary of connector capabilities from process files where:

• process kind is CALLABLE_SUB
• start element type is CallSubStart
• tags contains connector

The summary must be grouped by what a workflow team can actually do with the callable subs.

Inputs

• Optional file glob argument. Default: ./**/*.p.json
• Optional prose output file path. Default: stdout

Procedure

Before running, check the current OS. If on Windows, git bash or WSL is recommended to use for best compatibility.

1. Run the prose summary script (it reuses the JSON extractor internally):
   • bash ./.github/skills/callable-sub-summary/scripts/list-callable-sub-starts-json.sh
2. For file output, run:
   • bash ./.github/skills/callable-sub-summary/scripts/list-callable-sub-starts-json.sh './**/*.p.json' docs/callable-sub-capabilities-summary.txt
3. Inspect the generated callable-sub capability output and summarize it in marketing language.

Output Rules

When asked for a summary such as "give me a marketing summary" or "what features can I expect", produce this format:

• Intro sentence: one concise value statement.
• Section heading: What you get
• 4-6 bullets, each describing concrete workflow capabilities (prefer grouping by domain when present: calendar, mail, files, chat, todo).
• Section heading: At a glance
• 3-5 bullets with compact proof points such as callable count, domain coverage, read/write mix, typed payload usage, connector-tag discoverability.
• Section heading: Expected developer experience
• 2-4 bullets focused on developer outcomes (integration speed, reusable orchestration patterns, fit for real workflow use cases).
• Optional final sentence: clearly visible coverage limits only (for example input-only or result-only starts).

Additional requirements:

• Keep the response concise and scannable (typically 10-16 lines).
• Focus on signatures and typed input/result params as proof of capability.
• Use marketing language, but keep claims grounded in detected starts only.
• Prefer action-oriented phrasing that explains what teams can build.
• Do not paste raw JSON in the final summary.