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.
npx agentmods add skills/koroqe/opos/schedule-processnpx skills add Koroqe/OPOS --skill schedule-processgit clone --depth 1 https://github.com/Koroqe/OPOSWrote 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/skills/koroqe/opos/schedule-process)<a href="https://agentmods.dev/skills/koroqe/opos/schedule-process"><img src="https://agentmods.dev/badge/skills/koroqe/opos/schedule-process.svg" alt="Measured on agentmods" 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 | $0.00046 | $0.05076 |
| Opus 5 | $0.00023 | $0.02538 |
| Sonnet 5 | $0.00009 | $0.01015 |
| Haiku 4.5 | $0.00005 | $0.00508 |
Grade A, and why
schedule-process 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 4d 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 — 142 lines — stays where its author put it; the contents beside it link to each section on GitHub.
schedule-process
When to use
When a PROCESS.md has the 4 scheduling frontmatter fields set (schedule, runtime, non_interactive, authority — see ui/scheduling.py) and the user wants to activate the schedule. Invoke as /schedule-process <process-name>.
This skill is a wrapper around Claude Code's built-in CronCreate tool. It does not implement cron itself; it composes a routine prompt + authority prelude, validates the intent against the in-repo declaration, calls CronCreate, and caches the resulting routine id locally for fast lookup by sibling skills (unschedule-process, list-scheduled-processes).
Prerequisite: the user must be logged into Claude Code (claude login). The built-in cron tools authenticate against the user's Claude Code subscription. No additional OPOS setup is required.
Persistence contract (v0.9.1 — surface this to the user at every registration)
CronCreate's persistence varies by Claude Code build. In current builds registrations are session-scoped: jobs live in the running session's memory, are not written to disk, die when the session exits, and recurring jobs auto-expire after 7 days. A "registered" process therefore silently stops firing when the founder's session ends — worse than RISKS Risk 20's silent-failure case, because even the registration disappears.
Operating rules that follow:
- Step 8a below is mandatory: after
CronCreatereturns, inspect its response for the persistence markers (e.g. "session-only", "auto-expires") and print the contract to the user verbatim. Never let a registration look durable when the tool said otherwise. - Session-scoped mode → re-arm per session. Re-running
/schedule-process <name>is idempotent (step 7), so the standing remedy is simply to re-run it in each new session. The recommended self-healing setup is aSessionStarthook in the consumer's.claude/settings.jsonthat emits a reminder when.claude/scheduled-processes.jsonlists processes (the session then re-registers silently — the original registration already supplied the human authorization for the same unchanged declaration; any DIFFERENT cron/authority still goes through step 7's confirm):"hooks": {"SessionStart": [{"hooks": [{"type": "command", "command": "python3 -c \"import json,os; p='.claude/scheduled-processes.json'; rows=json.load(open(p)) if os.path.exists(p) else []; print('[opos-scheduler] ' + str(len(rows)) + ' scheduled process(es) declared (' + ', '.join(r['process_name'] for r in rows) + '). CronCreate registrations are session-scoped: re-run /schedule-process for each to re-arm this session (idempotent; previously authorized).') if rows else None\""}]}]} - Durable path A —
runtime: cloud(SHIPPED v0.14; RECOMMENDED for most consumers). Each fire spawns an isolated Claude Code session in Anthropic's cloud with its own fresh checkout of the repo, on the declared cron. Billed to the operator's Claude subscription — no Anthropic API key and no separate API billing. Prerequisite: the operator's GitHub account is connected to claude.ai (the Claude GitHub App, https://claude.ai/code/onboarding?magic=github-app-setup) with access to the repo — without it, registration fails with "Connect your GitHub account before saving a routine that uses a GitHub repository." The cloud session starts with ZERO context: the composed prompt must be self-contained (see step 4b). Runs are visible at https://claude.ai/code/routines with per-run logs — that page is the liveness signal for this runtime.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 4d ago First seen · 142 lines · 46 tokens per session scan A 81008bd85ec6
schedule-process is a skill published in the GitHub repository Koroqe/OPOS (2 stars, last pushed 2d ago), licensed MIT. It adds 46 tokens to every session and 5,076 once invoked, about $0.0002 per session on Opus 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 skills, from other repositories
agent-self-scheduling
Schedule AI agent runs with cron, loops, or external clocks while avoiding unsafe tight autonomous timers.
c-cron
Manage cron jobs and macOS launchctl services — schedule recurring tasks, start/stop services using lunchy-go.
convex-cron-jobs
Scheduled function patterns for background tasks including interval scheduling, cron expressions, job monitoring, retry strategies, and best practices for long-running tasks.
convex-cron-jobs
Scheduled function patterns for background tasks including interval scheduling, cron expressions, job monitoring, retry strategies, and best practices for long-running tasks.
agent-self-scheduling
Schedule AI agent runs with cron, loops, or external clocks while avoiding unsafe tight autonomous timers.
cron-doctor
Diagnose and validate cron expressions before they ship. Catches the five silent death-traps: impossible dates that never fire, OR-semantics that fire too often, midnight spikes, uneven step drift, and leap-year February 29.