Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/coltonbearden/carrelWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/coltonbearden/carrel/watch-folder)<a href="https://agentmods.dev/commands/coltonbearden/carrel/watch-folder"><img src="https://agentmods.dev/badge/commands/coltonbearden/carrel/watch-folder/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/coltonbearden/carrel/watch-folder"><img src="https://agentmods.dev/badge/commands/coltonbearden/carrel/watch-folder.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00020 | $0.01215 |
| Opus 5.5 | $0.00008 | $0.00486 |
| Sonnet 5.5 | $0.00004 | $0.00243 |
| Haiku 4.5 | $0.00002 | $0.00121 |
Grade A, and why
watch-folder scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 1mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 77 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Set up a folder watch for: $ARGUMENTS
Run the carrel CLI via Bash. Map the user's request onto the real flags in the --help block below (regenerated from the CLI by scripts/sync_plugins.py; if the installed carrel watch --help differs, trust the installed version — never invent flags):
Usage: carrel watch [OPTIONS] DIRECTORY
Watch DIRECTORY and run shell actions on file events.
Events for files an action is currently producing are suppressed via an in-flight set plus an
output-name heuristic (outputs whose name starts with the source file's stem); other action
outputs written into the watched directory WILL re-trigger — write outputs elsewhere or use --glob
to narrow matches. --stable waits for a file to stop growing, --existing processes what is already
there, --poll works where inotify does not (/mnt/c, shares), --done-dir/--error-dir file sources
away after their actions, --log keeps a JSON trail. Ctrl-C exits cleanly.
--done-dir/--error-dir refuse to start (exit 2) when they would move files git is tracking;
--allow-tracked overrides. Actions themselves are never guarded — what a --run command does is the
user's business.
Options:
--on EVENTS Comma-separated events to react to: created, modified, deleted,
moved, existing. [default: created,modified]
--glob PATTERN Only react to file names matching this glob (e.g. '*.pdf').
--run CMD Shell action to run per event; repeatable, runs in order. {path},
{name}, {stem}, {ext}, {dir} are substituted (shell-quoted).
[required]
--recursive Watch subdirectories too.
--existing Queue the files already in DIRECTORY at start (event 'existing').
--stable SECS Act only once a file's size and mtime have not changed for SECS
(scanners, big copies). [x>0]
--stable-timeout SECS With --stable: give up waiting and act after SECS regardless.
[x>0]
--poll Poll instead of inotify (needed on /mnt/c, network shares, some
containers).
--poll-interval SECS With --poll: how often to scan. [default: 1.0; x>0]
--done-dir DIRECTORY Move each source here after its actions all succeed.
--error-dir DIRECTORY Move each source here after an action fails.
--log FILE Append one JSON record per action (and per move) to FILE.
--allow-tracked With --done-dir/--error-dir: move files even when git tracks them.
--force Deprecated spelling of --allow-tracked; warns when it bypasses the
guard.
--print-service [systemd|schtasks]
Print a service definition that runs this exact watch at login,
then exit.
--debounce MS Coalesce events per path within this window. [default: 500; x>=0]
--once Exit after the first triggered action batch.
--timeout SECS Hard stop after SECS seconds. [x>0]
--action-timeout SECS Kill an action that runs longer than SECS (logged as rc=124).
[default: 300.0; x>0]
--json-lines Log one JSON object per action to stdout instead of human lines
(--json implies this).
--json Machine-readable JSON output.
--help Show this message and exit.
--on: comma-separated events (created, modified, deleted, moved); defaultcreated,modified.--glob: only react to matching filenames.--run CMD(required; repeatable, runs in order): the action;{path},{name}and{dir}are substituted (shell-quoted for you).--once: exit after the first triggered action batch — use for demos/tests.--timeout SECS: hard stop.--action-timeout SECS: kill a runaway action (logged as rc=124).--debounce MS(default 500): coalesce editor save-storms.--json-lines: machine-readable event log on stdout.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 1mo ago Changed · +2 lines 26ad8d8477a0
- 1mo ago Changed · +23 lines 51b60d680037
- 1mo ago Changed · +25 lines f58098e52631
- 1mo ago Changed d4c6990de9f2
- 1mo ago First seen · 27 lines · 20 tokens per session scan A 6819fd6bb188
watch-folder is a command published in the GitHub repository coltonbearden/carrel (2 stars, last pushed 2d ago), licensed MIT. It adds 20 tokens to every session and 1,215 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
review-resume
Comprehensive PM resume review against 10 best practices — structure, impact metrics, keywords, and actionable feedback.
follow-up
Review and act on pending reminders. Bare invocation shows all reminders as an indexed list; cron-fired invocation (/follow-up due ) surfaces the specific reminder and waits for a response; /follow-up review lists stale project follow-up entries for per-item disposition (extend/close/promote/skip). Transport-aware…
stop
Gracefully close the Vibe Squad tmux session from inside Chrono.
amp-download
Command "amp-download" from agentmessaging/claude-plugin, covering /amp-download, usage, arguments, options and examples.
todo.template
This prompt was authored for Claude-style slash workflows. In Codex runtime, adapt tool calls as follows.
add-skill
Add a new Skill to the current plugin with proper structure.