agentleFS
jonathanmalkinjonathanmalkin/jules/.claude/skills/reply-x/SKILL.md

reply-x

Skill65 starsChanged 7 months agoMITView on GitHub

Find posts where @[your-handle] tags @builtwithjules, draft replies, post approved ones as @builtwithjules.

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. About 0 are common, 6 are distinctive, and 8 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 8

    Very specific

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

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

    Read the tracking file: Documents/Content-Pipeline/X-Mentions-Jules.md

    line 16
  2. Distinctive:

    If mentions doesn't surface quote tweets, fall back to raw API:

    line 41
  3. Distinctive:

    If no new results after filtering, say so and stop.

    line 57
  4. Very specific:

    Do not assume "not tagged means blocked." If the post allows broad replies, Jules may be able to reply even when not explicitly tagged.

    line 69
  5. Distinctive:

    Don't draft a reply for truly ineligible tweets.

    line 72
  6. Distinctive:

    Keep under 280 characters when possible

    line 74
  7. Very specific:

    Add all new tweets to the tracking file immediately with presented: yes, action: pending.

    line 85
  8. Distinctive:

    Wait for [Your Name]'s approval before posting anything.

    line 87
  9. Distinctive:

    After [Your Name] responds with approve/skip decisions:

    line 99
  10. Very specific:

    Update each tweet's Action and Reply ID in the tracking file

    line 101
  11. Very specific:

    Update Last seen tweet ID to the highest tweet ID seen this run

    line 102
  12. Very specific:

    Update Last processed timestamp to now (CT)

    line 103
  13. Very specific:

    If posting fails with 401: run xurl auth oauth2 outside the Claude session, log in as @builtwithjules in the browser.

    line 117
  14. Very specific:

    If posting fails with 403 "not engaged" or equivalent restrictions after the eligibility check: the post is effectively closed to Jules.

    line 120

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “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

Replies

Find posts where @[your-handle] tags @builtwithjules, draft a reply for each, post after approval.

Voice: Jules — warm, direct, opinionated. Practitioner, not guru. Specific, not generic.

Step 1: Load State

Read the tracking file: Documents/Content-Pipeline/X-Mentions-Jules.md

Extract:

  • Last seen tweet ID — for filtering (not used in query — see note below)
  • Existing tweet IDs — to avoid re-presenting anything already shown

Step 2: Search

2a. Text search

xurl search "@builtwithjules from:[your-handle]" -n 20 -u builtwithjules

Filter results against the tracking file (skip any tweet ID already present).

2b. Mentions fetch (catches quote tweets)

Text search misses quote reposts where @builtwithjules appears in the quoted tweet but not in the quoting text. To catch these, fetch Jules's mentions directly:

xurl mentions -n 20 -u builtwithjules

If mentions doesn't surface quote tweets, fall back to raw API:

xurl /2/users/by/username/[your-handle] -u builtwithjules  # get user_id
xurl "/2/users/{user_id}/tweets?expansions=referenced_tweets.id&max_results=20" -u builtwithjules

Skip any tweet ID already in the tracking file.

2c. Merge and deduplicate

Combine results from 2a and 2b. Deduplicate by tweet ID.

Reply + standalone pairs: [Your Name] often posts the same content as both a reply and a standalone/quote tweet. When detected (same text or near-identical), group them and present as a single item. Default to replying to the conversational thread version.

If no new results after filtering, say so and stop. Do NOT update "Last processed" timestamp.

Step 3: Draft

For each new tweet, draft a reply:

  • Check reply eligibility first. Fetch the target post and inspect reply_settings:
xurl "/2/tweets/{tweet_id}?tweet.fields=reply_settings,conversation_id,author_id,created_at,public_metrics"

Do not assume "not tagged means blocked." If the post allows broad replies, Jules may be able to reply even when not explicitly tagged. If the post is limited (followers/mentioned users) and Jules is not eligible, then flag it: "Jules isn't eligible to reply here — tag Jules first or skip." Don't draft a reply for truly ineligible tweets.

  • Be specific and useful — no generic "great point!" responses
  • Keep under 280 characters when possible
  • Apply outbound sanitization (no sensitive data)

Present each for approval:

**Original** (@[your-handle]): [tweet text]
**Reply:** [draft]
**Action:** approve / edit / skip

Add all new tweets to the tracking file immediately with presented: yes, action: pending.

Wait for [Your Name]'s approval before posting anything.

Step 4: Post

For each approved reply:

xurl reply [TWEET_ID] "[reply text]" -u builtwithjules

Step 5: Update State

After [Your Name] responds with approve/skip decisions:

  1. Update each tweet's Action and Reply ID in the tracking file
  2. Update Last seen tweet ID to the highest tweet ID seen this run
  3. Update Last processed timestamp to now (CT)

Report: N new found, N posted, N skipped.


Auth Notes

Posts as @builtwithjules via -u builtwithjules using OAuth2 credentials stored in ~/.xurl.

xurl has two OAuth2 users under the [your-handle] app:

  • builtwithjules — used by /reply-x (-u builtwithjules)
  • [your-handle] — used by /watch-contacts (default, no flag needed)

If posting fails with 401: run xurl auth oauth2 outside the Claude session, log in as @builtwithjules in the browser.

If posting fails with 403 "not engaged" or equivalent restrictions after the eligibility check: the post is effectively closed to Jules. Ask [Your Name] to tag @builtwithjules on the tweet first, then retry.

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.