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 skills add softspark/ai-toolkit --skill repeatgit clone --depth 1 https://github.com/softspark/ai-toolkitWrote 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/softspark/ai-toolkit/repeat)<a href="https://agentmods.dev/skills/softspark/ai-toolkit/repeat"><img src="https://agentmods.dev/badge/skills/softspark/ai-toolkit/repeat/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/skills/softspark/ai-toolkit/repeat"><img src="https://agentmods.dev/badge/skills/softspark/ai-toolkit/repeat.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00033 | $0.01228 |
| Opus 5.5 | $0.00013 | $0.00491 |
| Sonnet 5.5 | $0.00007 | $0.00246 |
| Haiku 4.5 | $0.00003 | $0.00123 |
Grade A, and why
repeat 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 — 150 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/repeat - Autonomous Recurring Execution
$ARGUMENTS
Argument Parsing
Parse the user input into three components:
- Interval — duration between iterations (default:
10m)- Accepts:
1m,5m,10m,30m,1h - Minimum:
1m(enforced, reject anything lower)
- Accepts:
- Command or prompt — the slash command or freeform prompt to execute each iteration
- Flags:
--iterations N— maximum number of iterations (default:5)
Examples of parsed input:
/repeat 5m /test-> interval=5m, command=/test, iterations=5/repeat 10m "check deploy status"-> interval=10m, command="check deploy status", iterations=5/repeat --iterations 3 /review-> interval=10m, command=/review, iterations=3/repeat 2m --iterations 10 /lint-> interval=2m, command=/lint, iterations=10
Safety Controls
These limits are constitutionally mandated (Article I, Section 4). Violating them is not permitted.
Max Iterations
- Default: 5 iterations
- Configurable via
--iterations N - Hard ceiling: the loop MUST terminate after N iterations regardless of outcome
Circuit Breaker
- Track consecutive failures (non-zero exit code, error output, exceptions)
- 3 consecutive failures -> immediate hard stop
- Reset the failure counter on any successful iteration
Minimum Interval
- Never execute faster than 1 minute between iterations
- If user requests
<1m, override to1mand warn
Exit Detection
After each iteration, inspect the output for completion signals:
- Exit code
0combined with success markers - Output contains:
DONE,COMPLETE,ALL PASS,ALL TESTS PASSED,SUCCESS - Case-insensitive matching
- If detected: stop the loop early and report success
Execution Protocol
For iteration = 1 to max_iterations:
1. Log iteration start (timestamp, iteration number)
2. Execute the command or prompt
3. Capture output and exit status
4. Check exit detection:
- If completion marker found -> STOP, report success
5. Check circuit breaker:
- If failure: increment consecutive_failures
- If success: reset consecutive_failures to 0
- If consecutive_failures >= 3 -> HARD STOP, report failure
6. Log iteration result to stats
7. If not last iteration: wait for interval duration
8. Repeat
After loop ends:
- Print summary: total iterations, pass/fail counts, elapsed time
- Final status: COMPLETED | EARLY_SUCCESS | CIRCUIT_BREAKER | MAX_ITERATIONS
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 First seen · 150 lines · 33 tokens per session scan A c934cb1a7805
repeat is a skill published in the GitHub repository softspark/ai-toolkit (181 stars, last pushed today), licensed Apache-2.0. It adds 33 tokens to every session and 1,228 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 2026-09-03.
Other skills, from other repositories
work-autonomous
Abbruchbedingung für autonome Loops: So weit wie möglich selbständig weiterarbeiten UND einen Loop erst beenden, wenn belegt ist, dass keine autonom ausführbare Aufgabe mehr vorliegt. Der bloße Eindruck "nichts mehr zu tun" reicht NICHT — er löst eine vierstufige Prüfkette aus (think/decide, CONTROL/DECISIONS…
rotation-check
Standard-Gerüst für rotierende Pipeline-Checks: Pro Lauf genau ein Ziel aus einer Menge (Projekte, Ordner, Repos) wählen — bevorzugt das am längsten ungeprüfte —, den Check durchführen, Ergebnis in einer Check-Registry und einem Verlaufslog festhalten. Nutze diesen Skill, wenn ein wiederkehrender Check über viele…
kalender
Calendar skill with user-adaptive backend selection (Flag 3). Default: local SQLite store. Optional: Google Calendar MCP, Routinika or UpToday as backend — controlled via assist/prefs.json. Without preference the LLM asks the user interactively.
bewerbungsexperte
Spezialist fuer den gesamten Bewerbungsprozess. Analysiert Stellenanzeigen, optimiert Profile (LinkedIn/CV) und generiert massgeschneiderte Anschreiben. Generiert ASCII-Lebenslaeufe aus einer SQLite-Datenbank und Ordnerstruktur. cvgenerator.py ist standalone portiert -- keine BACH-Runtime noetig.
folder-flattening
Verschachtelte Ordnerhierarchien in flache, maschinenlesbare Strukturen umbauen. Bash-basiert mit intelligenter Merge-Logik.
batch-file-ops
Batch-Dateioperationen (Löschen, Verschieben, Kopieren, Auflisten) mit Glob-Mustern. CLI-Werkzeug für effiziente Dateisystem-Operationen. Ohne Abhängigkeiten.