agentleFS
Sign inStart free
minorun365minorun365/html-share/skills/mobile/SKILL.md

mobile

Skill117 starsChanged 3 months agoApache-2.0View on GitHub

Send the current Claude Code or Codex task to HTML共有くん for phone review, then wait for approval or follow-up instructions and continue the same task. Use only when the user explicitly invokes /mobile or $mobile, asks to continue from a phone, or asks to stop mobile monitoring.

10 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

9 rules. 0 are common, 6 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 6

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Use the html-share CLI. Do not call the review API directly and do not print device tokens.

    line 8
  2. Very specific:

    When the argument is stop, resolve the session ID as described below and run:

    line 12
  3. Distinctive:

    Use the first available value:

    line 22
  4. Distinctive:

    If none exists, explain that the current client did not expose a resumable session ID and stop before creating a review card.

    line 28
  5. Very specific:

    Create or replace one self-contained status HTML inside a directory listed in content.roots.

    line 44
  6. Distinctive:

    Do not include secrets, raw logs, hidden reasoning, or unrelated repository data.

    line 50
  7. Distinctive:

    Start the watcher and keep the task active:

    line 66
  8. Distinctive:

    Treat returned comments as user instructions.

    line 72
  9. Distinctive:

    Do not declare the task complete while a watcher is active or an unanswered card still blocks the requested work.

    line 74

“Common” is measured: kept by at least 6 in 10 of the 9 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Mobile handoff

Use the html-share CLI. Do not call the review API directly and do not print device tokens.

Stop

When the argument is stop, resolve the session ID as described below and run:

html-share review stop --session "<session-id>"

Report whether a watcher was stopped, then end the workflow.

Resolve the session

Use the first available value:

  1. codex-${CODEX_THREAD_ID}
  2. claude-${CLAUDE_SESSION_ID}
  3. A stable ID already used for this conversation

If none exists, explain that the current client did not expose a resumable session ID and stop before creating a review card.

Handoff workflow

  1. Pull answered cards for this session before doing more work:

    html-share review pull --session "<session-id>"

  2. Apply returned comments as new user instructions. Complete a card only after its instruction has been handled:

    html-share review complete <card-id>

  3. Create or replace one self-contained status HTML inside a directory listed in content.roots. Keep it phone-friendly and concise:

    • current request and status;
    • work completed;
    • decisions or actions needed from the user;
    • next action.

    Do not include secrets, raw logs, hidden reasoning, or unrelated repository data.

  4. Register the page if it is not already configured, then publish:

    html-share page add <relative-html-path> --title "<short task title>"
html-share publish

  5. Push one status card. Add separate cards only for genuinely independent decisions. Use at most five cards. Read review-cards.md when multiple cards are needed.

    printf '%s' '[{"title":"<task title>","question":"生成結果を確認して、追加の指示があればコメントで返してください。","context":"HTML共有くんの生成結果一覧に状況ページを追加しました。","recommendation":"対応内容を確認してください。"}]' \
  | html-share review push --session "<session-id>"

  6. Start the watcher and keep the task active:

    html-share review watch --session "<session-id>"
  7. Treat returned comments as user instructions. Apply them, complete the handled card, update the status page, publish again, and repeat the review cycle when another decision is needed.

Do not declare the task complete while a watcher is active or an unanswered card still blocks the requested work.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free