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 Bono12138/bonobox --skill reality-groundinggit clone --depth 1 https://github.com/Bono12138/bonoboxWrote 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/bono12138/bonobox/reality-grounding)<a href="https://agentmods.dev/skills/bono12138/bonobox/reality-grounding"><img src="https://agentmods.dev/badge/skills/bono12138/bonobox/reality-grounding/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/bono12138/bonobox/reality-grounding"><img src="https://agentmods.dev/badge/skills/bono12138/bonobox/reality-grounding.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 | $0.00083 | $0.01984 |
| Opus 5 | $0.00042 | $0.00992 |
| Sonnet 5 | $0.00017 | $0.00397 |
| Haiku 4.5 | $0.00008 | $0.00198 |
Grade A, and why
reality-grounding 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 7d 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 — 209 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Reality Grounding
Find enough reality to make the next decision without inventing an ideal organisation. This Skill is not a questionnaire and not a one-turn advice generator. It owns an active loop of inquiry, action, observed feedback and revision until the user confirms a real stopping point.
Capability version: 0.4.0.
Read references/active-inquiry.md when material facts are missing, the user
asks for proactive investigation, or the available source should be chosen.
Read references/reality-record-contract.md when the task needs a durable case
record or will feed $reality-strategy. Use
scripts/validate_reality_record.py before treating a substantial reality
record as ready for strategy. Use references/evaluation-cases.md when testing
or changing this Skill's routing and inquiry behaviour.
Lock the decision first
Identify the decision the user must make, the desired result, unacceptable downside, time limit, and which new fact could change the action. If there is no real decision yet, clarify the problem before collecting broad background.
Separate:
- verified fact;
- direct user observation or report;
- formal rule;
- observed practice;
- supported inference;
- working hypothesis;
- unknown.
A document proves what it says, not that people follow it. A repeated Agent claim does not turn a hypothesis into a fact.
Before substantial advice, decide whether the current information can actually separate the available routes. If a material unknown changes the actor, first step, feasibility or social risk, ask first. A short preview of the branches may explain the question; it must not become a premature catalogue of advice.
Recover the original problem before accepting the proposed method
The user's account may mix an original need with a proposed mechanism. Do not assume that solving the named mechanism solves the real problem. Write both:
- A — original result: what must become true in the real world;
- B — proposed route: the process, system, permission, meeting, tool or other mechanism currently being discussed.
What ships with it
5 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.
- 7d ago First seen · 209 lines · 83 tokens per session scan A a14b07e32b81
reality-grounding is a skill published in the GitHub repository Bono12138/bonobox (5 stars, last pushed 6d ago), licensed MIT. It adds 83 tokens to every session and 1,984 once invoked, about $0.0004 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 2026-09-05.
Other skills, from other repositories
ai-ml-development
AI and machine learning development with PyTorch, TensorFlow, and LLM integration. Use when building ML models, training pipelines, fine-tuning LLMs, or implementing AI features.
python-logging-best-practices
Python logging with loguru, structlog, and orjson. TRIGGERS - loguru, structlog, structured logging.
python-memory-safe-scripts
Memory-safe Python script patterns for long-running processes under systemd MemoryMax constraints. Covers allocator purge (mimalloc/glibc malloctrim), HTTP response lifecycle, DataFrame cleanup, thread-local connection reuse, and periodic GC cadence. Battle-tested through 5 OOM optimization cycles on production GPU…
pypi-doppler
LOCAL-ONLY PyPI publishing with Doppler credentials. TRIGGERS - publish to PyPI, pypi upload, local publish. NEVER use in CI/CD.
python-workspace
Python workspace for MQL5 integration. TRIGGERS - MetaTrader 5 Python, mt5 package, MQL5-Python setup.
python-performance
Profile and optimize Python code using cProfile, memory profilers, and performance best practices. Use when debugging slow Python code, optimizing bottlenecks, or improving application performance.