ss

Load screenshot images or other files from Dropbox screenshots directory. Use when user invokes /ss directly. NEVER manually list/read files from the screenshots dir — this skill handles Dropbox sync delays, freshness checks, retry logic that manual reads miss. Supports: /ss 2 (latest 2), /ss latest3, /ss filename.png (exact/substring), /ss full-path. Also non-image files (e.g., /ss pattern-4-variations.html).

6 stars

Best use case

ss is best used when you need a repeatable AI agent workflow instead of a one-off prompt.

Load screenshot images or other files from Dropbox screenshots directory. Use when user invokes /ss directly. NEVER manually list/read files from the screenshots dir — this skill handles Dropbox sync delays, freshness checks, retry logic that manual reads miss. Supports: /ss 2 (latest 2), /ss latest3, /ss filename.png (exact/substring), /ss full-path. Also non-image files (e.g., /ss pattern-4-variations.html).

Teams using ss should expect a more consistent output, faster repeated execution, less prompt rewriting.

When to use this skill

  • You want a reusable workflow that can be run more than once with consistent structure.

When not to use this skill

  • You only need a quick one-off answer and do not need a reusable workflow.
  • You cannot install or maintain the underlying files, dependencies, or repository context.

Installation

Claude Code / Cursor / Codex

$curl -o ~/.claude/skills/ss/SKILL.md --create-dirs "https://raw.githubusercontent.com/Takazudo/claude-resources/main/skills/ss/SKILL.md"

Manual Installation

  1. Download SKILL.md from GitHub
  2. Place it in .claude/skills/ss/SKILL.md inside your project
  3. Restart your AI agent — it will auto-discover the skill

How ss Compares

Feature / AgentssStandard Approach
Platform SupportNot specifiedLimited / Varies
Context Awareness High Baseline
Installation ComplexityUnknownN/A

Frequently Asked Questions

What does this skill do?

Load screenshot images or other files from Dropbox screenshots directory. Use when user invokes /ss directly. NEVER manually list/read files from the screenshots dir — this skill handles Dropbox sync delays, freshness checks, retry logic that manual reads miss. Supports: /ss 2 (latest 2), /ss latest3, /ss filename.png (exact/substring), /ss full-path. Also non-image files (e.g., /ss pattern-4-variations.html).

Where can I find the source code?

You can find the source code on GitHub using the link provided at the top of the page.

SKILL.md Source

# Screenshot Loader

Load screenshot images or other files from the Dropbox screenshots directory and present them in the conversation. Also supports non-image files (HTML, text, etc.) that the user shares via this directory.

## Resolve screenshots directory

Use `$DROPBOX_SCREENSHOTS_DIR` env var (set in `.zshrc` for both macOS and WSL2).

## Invocation timestamp

The following epoch was captured at the moment this command was invoked:

**Invocation epoch:** !`date +%s`

Use this as a cutoff for "Latest N" mode — only consider image files whose modification time is **<= this epoch**. This prevents picking up screenshots that appeared after the user typed `/ss`.

## Parse arguments

`$ARGUMENTS` determines which files to load:

| Pattern | Meaning | Example |
| --- | --- | --- |
| Bare number (`2`, `3`) | Latest N images at invocation time | `/ss 2` |
| `latestN` | Latest N images at invocation time | `/ss latest3` |
| Full path starting with `/` | Exact file | `/ss '/Users/.../file.png'` |
| Non-image filename (`.html`, `.txt`, etc.) | Read a non-image file from screenshots dir | `/ss pattern-4-variations.html` |
| Other string (exact) | Exact filename in screenshots dir | `/ss Screenshot 2026-03-14 at 3.27.17.png` |
| Other string (partial) | Substring search in filenames | `/ss foo-bar-moo.png` |
| (empty) | Latest 1 image | `/ss` |

**Default:** When `$ARGUMENTS` is empty or blank, treat it as `1` (load the latest single screenshot). This is the most common use case.

**Non-image files:** When the argument has a non-image extension (e.g., `.html`, `.txt`, `.json`, `.css`, `.js`, `.svg`, `.md`), look for that file in `$DROPBOX_SCREENSHOTS_DIR` and read it as a text file using the Read tool. This is for when the user shares files (not just screenshots) via the Dropbox screenshots directory.

## Find files

### Latest N images (with timestamp cutoff)

List all image files (png, jpg, jpeg, gif, webp, tiff) in the screenshots directory, filter to only those with modification time <= the invocation epoch, sort by modification time descending, take the first N.

```bash
CUTOFF=<invocation_epoch>
ls -t "$DROPBOX_SCREENSHOTS_DIR"/*.{png,jpg,jpeg,gif,webp,tiff} 2>/dev/null | while IFS= read -r f; do
  [ "$(stat -f %m "$f")" -le "$CUTOFF" ] && echo "$f"
done | head -n $N
```

This ensures that even if new screenshots appear while Claude is processing, only files that existed at invocation time are selected.

### Non-image file

When the argument has a non-image extension (not `.png`, `.jpg`, `.jpeg`, `.gif`, `.webp`, `.tiff`):

1. Look for the exact file at `$DROPBOX_SCREENSHOTS_DIR/<filename>`
2. If not found, do a substring search across **all** files (not just images) in the directory
3. Read the file with the Read tool (it will be read as text, not presented visually)

### Specific filename (image)

1. First, try exact match: `$DROPBOX_SCREENSHOTS_DIR/<filename>`
2. If the exact file does not exist, perform a **substring search** across all image files in the screenshots directory:
- Strip the extension from the argument to get the search term (e.g., `foo-bar-moo.png` → `foo-bar-moo`)
- Search filenames (case-insensitive) that contain the search term as a substring
- Also try **plural/singular variants**: if the search term contains a word, try both its singular and plural forms (e.g., `screenshot` ↔ `screenshots`, `image` ↔ `images`)
- If multiple files match, pick the most recently modified one
- If no matches found, proceed to the "Wait for file" step using the exact path

```bash
# Substring search example
SEARCH="foo-bar-moo"
find "$DROPBOX_SCREENSHOTS_DIR" -maxdepth 1 -type f \( -iname "*.png" -o -iname "*.jpg" -o -iname "*.jpeg" -o -iname "*.gif" -o -iname "*.webp" -o -iname "*.tiff" \) -iname "*${SEARCH}*" -print0 | xargs -0 ls -t 2>/dev/null | head -n 1
```

### Full path

Use the path as-is.

## Wait for file — Dropbox sync delay handling

The user typically takes a screenshot and immediately runs `/ss`. Dropbox sync introduces a delay of a few seconds (sometimes longer), so the file may not exist locally yet. This section handles that proactively.

### Initial delay for "Latest N" mode

For "Latest N" mode (bare number, `latestN`, or empty argument), **always sleep 3 seconds before the first file lookup**. This gives Dropbox a moment to sync the new screenshot. This simple delay avoids the most common case where old files are returned because the new one hasn't arrived yet.

```bash
# Always do this before the first file listing in "Latest N" mode
sleep 3
```

### Freshness check for "Latest N" mode

After listing files, check whether the results are **fresh enough** relative to the invocation epoch. A screenshot the user just took should have a modification time within ~60 seconds of the invocation epoch.

**Freshness rule:** At least one of the N files must have a modification time within 120 seconds before the invocation epoch. If ALL files are older than 120 seconds before the invocation epoch, the user's new screenshot likely hasn't synced yet — retry.

```bash
CUTOFF=<invocation_epoch>
FRESHNESS_THRESHOLD=$((CUTOFF - 120))
# After getting the N files, check the newest one:
NEWEST_MTIME=$(stat -f %m "$NEWEST_FILE")
if [ "$NEWEST_MTIME" -lt "$FRESHNESS_THRESHOLD" ]; then
  # All results are stale — the new screenshot hasn't synced yet, retry
fi
```

### Retry loop

If the file is not found (specific filename mode) or the freshness check fails (Latest N mode), poll with retries:

1. **Poll every 5 seconds** for up to **2 minutes** (24 attempts)
2. For **specific filename mode**: check if the target file exists (`[ -f "$TARGET" ]`)
3. For **"Latest N" mode**: re-list files and re-check the freshness rule (at least one file within 120 seconds of invocation epoch)
4. If the file appears or freshness check passes during polling, proceed normally
5. If timeout is reached, report that the file was not found (or that only older files were found) and ask the user to confirm

```bash
# Specific filename mode
for i in $(seq 1 24); do
  [ -f "$TARGET" ] && break
  sleep 5
done

# Latest N mode (after initial 3-second delay already elapsed)
for i in $(seq 1 24); do
  # Re-run the file listing and freshness check
  FILES=$(ls -t "$DROPBOX_SCREENSHOTS_DIR"/*.{png,jpg,jpeg,gif,webp,tiff} 2>/dev/null | while IFS= read -r f; do
    [ "$(stat -f %m "$f")" -le "$CUTOFF" ] && echo "$f"
  done | head -n $N)
  NEWEST=$(echo "$FILES" | head -n 1)
  [ -n "$NEWEST" ] && [ "$(stat -f %m "$NEWEST")" -ge "$FRESHNESS_THRESHOLD" ] && break
  sleep 5
done
```

### Summary of the wait flow for "Latest N" mode

1. Sleep 3 seconds (initial delay)
2. List files, apply timestamp cutoff, take top N
3. Check freshness: is the newest file within 120 seconds of invocation epoch?
4. If fresh enough, proceed to present files
5. If stale, enter retry loop (poll every 5s, up to 2 minutes)
6. If timeout, present what was found but warn the user that these may be older files

## Present the files

Use the Read tool to read each file. The Read tool supports reading image files (PNG, JPG, etc.) visually, and text files as content.

After reading, briefly acknowledge which file(s) were loaded (filename only, not full path) and ask what the user would like to do with them.

## CRITICAL: Relevance check — never silently ignore loaded files

**After presenting files, you MUST treat them as the user's intended files.** The user just took these screenshots and ran `/ss` — they expect you to work with whatever was loaded.

**NEVER do this:**

- Load files, decide they "don't seem related," and ignore them
- Skip the screenshots and continue with the conversation as if `/ss` wasn't invoked
- Say "these don't appear to be relevant" and move on

**The user's screenshots ARE the context.** Even if the content seems unrelated to the prior conversation, the user may be introducing new context, switching topics, or showing something you don't yet understand.

### If files seem potentially mismatched (e.g., clearly old/stale screenshots)

Only if the files are **obviously stale** (e.g., timestamps are many hours old, content is clearly from a different session), follow this two-step recovery:

**Step 1: Wait and recheck (Dropbox sync delay)**

The user's actual screenshot may still be syncing. Wait 2 minutes, then re-list the directory for newer files:

```bash
# Wait for potential Dropbox sync delay
sleep 120

# Re-check for newly synced files (allow mtime up to 180s after invocation epoch)
LATE_CUTOFF=$((CUTOFF + 180))
ls -t "$DROPBOX_SCREENSHOTS_DIR"/*.{png,jpg,jpeg,gif,webp,tiff} 2>/dev/null | while IFS= read -r f; do
  MTIME=$(stat -f %m "$f")
  [ "$MTIME" -le "$LATE_CUTOFF" ] && [ "$MTIME" -gt "$CUTOFF" ] && echo "[NEW] $f"
done | head -n 3
```

If new files appeared, read and present them — these are likely the intended screenshots.

**Step 2: Ask the user**

If no new files appeared after waiting, present what you have and **ask the user to confirm**:

> "I loaded [filenames]. These screenshots appear to be from [timestamp]. Are these the ones you meant, or should I wait longer for a newer screenshot to sync?"

**Never skip both steps.** If you suspect the files are wrong, you MUST either wait-and-recheck OR ask the user. Silently ignoring the user's `/ss` invocation is never acceptable.

Related Skills

zudoesa-articlify

6
from Takazudo/claude-resources

Convert conversation context into an esa article via the zudoesa-writer subagent. ONLY invoke when the user explicitly asks — NEVER proactively propose. Triggers: 'write esa article', 'esa記事', 'esaに書いて', 'articlify for esa', or /zudoesa-articlify. Gathers context, creates a writing brief, delegates to the writer subagent.

zudoesa-apply-voice

6
from Takazudo/claude-resources

Apply Takazudo's esa writing voice and vocabulary rules to text. Use when: (1) User wants to write/rewrite text in Takazudo's esa style, (2) User says 'apply voice', 'esa voice', 'esa文体で', 'esa風に書いて', '文体を適用', (3) User provides text to transform to esa style. Reads writing-style.md and vocabulary-rule.md from takazudo-esa-writing repo and applies the rules.

zudocg-articlify

6
from Takazudo/claude-resources

Convert conversation context into a CodeGrid article via the zudocg-writer subagent. ONLY invoke when the user explicitly asks — NEVER proactively propose. Triggers: 'write codegrid article', 'CodeGrid記事', 'codegridに書いて', 'articlify for codegrid', or /zudocg-articlify. Gathers context, creates a writing brief, delegates to the writer subagent.

zudocg-apply-voice

6
from Takazudo/claude-resources

Apply Takazudo's CodeGrid writing voice and vocabulary rules to text. Use when: (1) User wants to write/rewrite text in Takazudo's CodeGrid style, (2) User says 'apply voice', 'codegrid voice', 'codegrid文体で', 'codegrid風に書いて', '文体を適用', (3) User provides text to transform to CodeGrid style. Reads writing-style.md and vocabulary-rule.md from takazudo-codegrid-writing repo and applies the rules.

zpaper-articlify

6
from Takazudo/claude-resources

Convert conversation context into a zpaper blog article via the zpaper-writer subagent. ONLY invoke when the user explicitly asks — NEVER proactively propose. Triggers: 'write zpaper article', 'zpaper記事', 'zpaperに書いて', 'articlify for zpaper', or /zpaper-articlify. Gathers context, creates a writing brief, delegates to the writer subagent.

zpaper-apply-voice

6
from Takazudo/claude-resources

Apply Takazudo's zpaper blog writing voice and vocabulary rules to text. Use when: (1) User wants to write/rewrite text in Takazudo's zpaper style, (2) User says 'apply voice', 'zpaper voice', 'zpaper文体で', 'zpaper風に書いて', 'ブログ文体を適用', (3) User provides text to transform to zpaper style. Reads writing-style.md and vocabulary-rule.md from the zpaper repo and applies the rules.

xlsx

6
from Takazudo/claude-resources

Spreadsheet creation, editing, and analysis. Use when working with .xlsx, .xlsm, .csv, .tsv files for: (1) Creating spreadsheets with formulas and formatting, (2) Reading or analyzing data, (3) Modifying existing spreadsheets while preserving formulas, (4) Data analysis and visualization, (5) Recalculating formulas.

x

6
from Takazudo/claude-resources

Facade for development workflows. Routes on two axes: plan-first vs implement-now (escalates to /big-plan -a when the request needs research / decomposition / has unclear scope — the appended -a makes the plan chain into implementation in-session), then single vs multi on the ready-to-build fast paths (/x-as-pr single-topic, /x-wt-teams multi-topic parallel). Use when: (1) User says '/x' followed by dev instructions, (2) User wants to start development without choosing the workflow skill, (3) User says 'dev', 'implement', or 'build' with a task. Default option: -v (verify-ui). Review-loop (-l) is opt-in — without -l the downstream skill runs a single /deep-review pass. Forwards -a (autonomy/auto-chain) and -m (merge at the end + cleanup + CI watch) through every route; auto-fix of raised findings (-f) and issue-raising (-ri) are downstream defaults, with -nf/--no-fix and -nori/--no-raise-issues as the forwarded opt-outs. -a and -m are orthogonal — full hands-off end-to-end is -a -m.

x-wt-teams

6
from Takazudo/claude-resources

Parallel multi-topic development using git worktrees, base branches, and Claude Code agent teams. Use when: (1) User wants to work on multiple related features in parallel, (2) User mentions 'worktree', 'base branch', 'parallel development', 'split into topics', or 'multi-topic'. FULLY AUTONOMOUS — creates worktrees, spawns teams, coordinates everything. Also supports Super-Epic child mode for [Epic] issues from /big-plan with '**Super-epic:** #N' markers (targets the super-epic base branch instead of main).

x-as-pr

6
from Takazudo/claude-resources

Start a development workflow as a draft PR. Creates a NEW branch from the current branch, empty start commit, draft PR targeting the current branch, then implements. ALWAYS creates a new branch by default — produces a nested PR-on-PR when the current branch already has one. Use when: (1) User says 'dev as pr', (2) User wants a PR-first workflow before coding, (3) User passes -s/--stay to reuse the current branch instead of nesting, (4) User passes a GitHub issue URL to implement, (5) User passes --make-issue/--issue to create an issue first. Logs progress via issue comments when an issue is linked.

watch-ci

6
from Takazudo/claude-resources

Watch GitHub PR CI checks in the background and notify on completion. Use when: (1) User wants to monitor CI/CD status, (2) User says 'watch CI', 'check CI', 'monitor checks', or 'wait for CI', (3) User wants to know when checks pass or fail. Runs a background gh polling shell loop (NOT a subagent — near-zero token cost), sends macOS notification on completion. Also handles merged PRs by watching the target branch CI.

w-update-wording-rule

6
from Takazudo/claude-resources

Add or update wording rules (表記ルール) in the w repo's vocabulary-rule.md files. Use when: (1) User says 'add wording rule', 'update wording rule', '表記ルール追加', (2) User wants to add a kanji/hiragana usage rule, (3) User provides a rule like 'X should be Y' with examples.