feat(domain-skills/x): X.com post domain skill

[miladnoo]

Summary

• domain-skills/x/x_post.py — post text + image(s) and/or video(s) to X.com
• domain-skills/x/post.md — full documentation
• SKILL.md — updated to list the new domain skill

Features

• Single or multiple images — auto-detected by extension, or use img: prefix
• Video with auto H.264 conversion — non-H.264 video is converted via ffmpeg before upload
• Mixed media — images + video in one post, sequential upload
• Explicit type override — img: or vid: prefix when extension isn't reliable
• JS click on Post button — coordinate click fails after media attachment (X re-renders button position)
• Polling loop — waits for aria-disabled != 'true' to handle slow video processing

Test Results

All three modes verified on X.com:

• Image only ✅
• Video only ✅
• Image + Video mixed ✅

🤖 Generated with Claude Code

---

Summary by cubic

Adds an X.com posting domain skill to publish text with image(s) and/or video via the browser harness. Includes domain-skills/x/post.md and updates SKILL.md to list the skill.

• New Features

  • Post text with images and/or video; supports multiple and mixed media with auto-detect and img:/vid: overrides.
  • Converts non-H.264 video to H.264/AAC (ffmpeg); uses JS click + polling until Post is enabled to handle video processing.

• Bug Fixes

  • Detects ffmpeg conversion failures (check=True) and avoids returning bad temp paths.
  • Always cleans up converted temp files, including early-return paths (wrapped in try/finally), with an atexit safety net.

^{Written for commit 6655b39. Summary will update on new commits.}

[browser-harness-review]

✅ Skill review passed

Reviewed 3 file(s) — no findings.

[cubic-dev-ai]

2 issues found across 3 files

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="domain-skills/x/x_post.py">

<violation number="1" location="domain-skills/x/x_post.py:66">
P2: Converted H.264 temp videos are never deleted after use, causing orphaned `.mp4` files to accumulate.</violation>

<violation number="2" location="domain-skills/x/x_post.py:70">
P1: ffmpeg failures are not detected; the temp output path is returned even when conversion fails.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review, or fix all with cubic.}

[cubic-dev-ai]

1 issue found across 1 file (changes from recent commits).

Prompt for AI agents (unresolved issues)

Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.


<file name="domain-skills/x/x_post.py">

<violation number="1" location="domain-skills/x/x_post.py:212">
P2: Temp video files can leak on early-return failure paths because cleanup only runs after a successful post.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review, or fix all with cubic.}

[qodo-ai-reviewer]

Hi, domain-skills/x/x_post.py inserts domain-skills/ into sys.path but imports helpers/admin from repo root, so running or importing this file commonly fails with ModuleNotFoundError.
This prevents the skill from executing in the default script/module usage shown in the docs.

Severity: action required | Category: correctness

How to fix: Insert repo root to sys.path

Agent prompt to fix - you can give this to your LLM of choice:

Issue description

domain-skills/x/x_post.py modifies sys.path incorrectly, causing from helpers import * / from admin import ensure_daemon to fail when the script is executed or imported in common ways.

Issue Context

• helpers.py and admin.py are located at the repository root.
• The skill file is located at domain-skills/x/x_post.py.

Fix Focus Areas

• domain-skills/x/x_post.py[32-35]

Proposed fix

• Change the inserted path to the repository root (e.g., SKILL_DIR.parents[2] or SKILL_DIR.parent.parent.parent) and preferably use resolve().
• Consider avoiding import * and importing only what is needed for clarity.

We noticed a couple of other issues in this PR as well - happy to share if helpful.

---

Found by Qodo. Free code review for open-source maintainers.

[qodo-ai-reviewer]

Hi, post_to_x passes media paths directly to upload_file, but helpers.upload_file expects absolute file paths for DOM.setFileInputFiles. Passing relative paths (common in CLI usage) will cause uploads to fail at runtime.

Severity: action required | Category: reliability

How to fix: Resolve media paths to absolute

Agent prompt to fix - you can give this to your LLM of choice:

Issue description

upload_file() requires absolute paths but x_post.py forwards raw user inputs.

Issue Context

Users often pass ./file.png or ../video.mp4.

Fix Focus Areas

• domain-skills/x/x_post.py[140-177]

Suggested approach

Before splitting into images/videos and before _ensure_h264, normalize each m["path"] via Path(path).expanduser().resolve() (and possibly validate existence) so upload_file() always receives an absolute path.

We noticed a couple of other issues in this PR as well - happy to share if helpful.

---

Found by Qodo code review. FYI, Qodo is free for open-source.