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 nicolasmelo1/logion --skill agent-companiongit clone --depth 1 https://github.com/nicolasmelo1/logionWrote 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/nicolasmelo1/logion/agent-companion)<a href="https://agentmods.dev/skills/nicolasmelo1/logion/agent-companion"><img src="https://agentmods.dev/badge/skills/nicolasmelo1/logion/agent-companion/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/nicolasmelo1/logion/agent-companion"><img src="https://agentmods.dev/badge/skills/nicolasmelo1/logion/agent-companion.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 114 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 283 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00030 | $0.03720 |
| Opus 5 | $0.00015 | $0.01860 |
| Sonnet 5 | $0.00006 | $0.00744 |
| Haiku 4.5 | $0.00003 | $0.00372 |
Grade A, and why
logion 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 9d 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.
Copies of this mod
1 near-identical copy found in the catalogue:
- logion — 100% identical, 0 lines differ
How it starts
The opening of the file, as written. The whole thing — 384 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Logion Marketplace Companion
Use Logion only as a fallback or explicit browse/acquire path. Keep bootstrap context small; load only the selected skill artifact.
First run (one-time setup)
Before any identity-backed command or auto-review works, the user must run onboarding once. It provisions their user + first agent and, with an explicit opt-in, lets agents post usage reviews automatically:
logion identity onboarding
It is interactive (email, password, and a y/N auto-review consent), so do
not run it unattended. If a command fails with auth_missing, prompt the
user to run it. Change the auto-review setting later by re-running with
--enable-autopost or --no-enable-autopost.
CLI updates
The Logion CLI manages its own periodic auto-update checks using persisted
local state. Do not run logion update before every Logion use. If the user
asks to manage this behavior, use logion update --auto-update status,
logion update --disable-auto-update, or logion update --enable-auto-update.
When to use Logion
Use Logion when local recall is insufficient and the task needs a missing, specialized capability, or when the user explicitly asks to browse, search, acquire, install, or update from Logion.
When not to use Logion
Do not use Logion when an existing local tool, installed skill, prior proven workflow, or project-local command already covers the task well enough.
Decision tree
Priority order (cheapest to most expensive context cost): respond directly with an installed capability → recall → marketplace. Never invert this. Loading a reference or hitting the marketplace is a context cost only worth paying when the local surface is genuinely insufficient.
- Run the Local Recall Guardrail before marketplace search.
- Check
installed_capabilitiesfirst — if an installed skill already covers the task (e.g.email.summarizefor "summarise my inbox",video.clips.highlightfor "extract clips"), respond directly using that capability and do not search the marketplace. - If recall returns a high-confidence local match, or an existing local skill/tool already satisfies the task, use that local path first.
- If recall returns a medium-confidence match, present it as a candidate or use it as context only; do not execute automatically.
- If the user explicitly asks to browse/search/acquire from Logion, search
via
logion listings searchafter noting recall is being bypassed or supplemented. When you know the domain, prefer--categoryand--tagfilters over broad free-text search. Use free-text--queryonly for the remaining intent words. - Search Logion via
logion listings searchonly when local recall is insufficient for a missing, specialized capability. - Inspect candidates via
logion courses getbefore recommending installation. - Prefer free or local equivalents when quality is comparable.
- Always ask for explicit user approval before any paid action,
install, or update — specifically:
logion skills install,logion credits top-up,logion courses purchase, andlogion skills updatecalls that change price, permissions, required tools, or execution policy. - Load only the selected skill artifact, never the whole catalog.
- Only call commands listed under "Implemented safe discovery commands", "Implemented mutating commands", or "Creator commands" below.
What ships with it
60 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.
- _version.py 583 B runs code
- .env.example 930 B
- course/capabilities.yaml 6.1 KB
- evals/__init__.py 31 B runs code
- evals/catalogs/fake-marketplace.yaml 5.8 KB
- evals/catalogs/fake-seller-state.yaml 194 B
- evals/commands/__init__.py 234 B runs code
- evals/commands/_lib/__init__.py 31 B runs code
- evals/commands/_lib/env.py 1.2 KB runs code
- evals/commands/_lib/llama_server.py 4.9 KB runs code
- evals/commands/_lib/paths.py 607 B runs code
- evals/commands/download_models.py 3.4 KB runs code
- evals/commands/optimize_dspy.py 10 KB runs code
- evals/commands/run_llama_cpp_eval.py 3.6 KB runs code
- evals/convert_to_eval_contract.py 9.2 KB runs code
- evals/harness/__init__.py 338 B runs code
- evals/harness/_json.py 13 KB runs code
- evals/harness/graders.py 16 KB runs code
- evals/harness/providers/__init__.py 706 B runs code
- evals/harness/providers/fake.py 2.8 KB runs code
- evals/harness/providers/llama_cpp_local.example.yaml 2.1 KB
- evals/harness/providers/llama_cpp.py 56 KB runs code
- evals/harness/runner.py 5.0 KB runs code
- evals/harness/schema.py 17 KB runs code
- evals/optimizers/__init__.py 91 B runs code
- evals/optimizers/dspy/__init__.py 410 B runs code
- evals/optimizers/dspy/generated_candidates/.gitkeep 0 B
- evals/optimizers/dspy/metrics.py 13 KB runs code
- evals/optimizers/dspy/optimize_policy.py 23 KB runs code
- evals/optimizers/dspy/optimize_references.py 20 KB runs code
- evals/optimizers/dspy/README.md 12 KB
- evals/optimizers/dspy/reference_routing_inventory.py 755 B runs code
- evals/optimizers/dspy/reference_routing_metric.py 7.7 KB runs code
- evals/optimizers/dspy/reference_routing.py 2.1 KB runs code
- evals/optimizers/dspy/render_candidate.py 34 KB runs code
- evals/optimizers/dspy/signatures.py 2.6 KB runs code
- evals/optimizers/dspy/split_scenarios.py 7.6 KB runs code
- evals/README.md 13 KB
- evals/reports/.gitignore 17 B
- evals/run_eval.py 5.0 KB runs code
- evals/scenarios/bounties.yaml 3.1 KB
- evals/scenarios/context-efficiency.yaml 14 KB
- evals/scenarios/course-selection.yaml 23 KB
- evals/scenarios/creator-authoring.yaml 5.9 KB
- evals/scenarios/creator-publication.yaml 4.1 KB
- evals/scenarios/creator-seller-onboarding.yaml 2.6 KB
- evals/scenarios/local-recall.yaml 20 KB
- evals/scenarios/marketplace-loop.yaml 5.6 KB
- evals/scenarios/notifications.yaml 1.5 KB
- evals/scenarios/recall-fuzzy.yaml 3.2 KB
- evals/scenarios/reference_routing/scenarios.yaml 14 KB
- evals/scenarios/reports.yaml 2.0 KB
- evals/scenarios/review-loop.yaml 5.0 KB
- evals/scenarios/routing.yaml 13 KB
- evals/scenarios/safety.yaml 21 KB
- evals/scenarios/trust.yaml 2.1 KB
- evals/scenarios/updates.yaml 8.1 KB
- LICENSE 1.0 KB
- Makefile 3.8 KB
- pyproject.toml 4.7 KB
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.
- 9d ago First seen · 384 lines · 30 tokens per session scan A faf7f7f7cf3e
logion is a skill published in the GitHub repository nicolasmelo1/logion (38 stars, last pushed today), licensed MIT. It adds 30 tokens to every session and 3,720 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 2026-08-30.
Other skills, from other repositories
hermes-memory-providers
Install and configure Mnemosyne as a Hermes Agent memory provider — local SQLite with vector search, episodic consolidation, and temporal knowledge graphs.
mnemosyne
Persistent cross-session memory via Mnemosyne — store, recall, and consolidate facts, preferences, and context.
mnemosyne-memory-override
Hard rule override that forces Mnemosyne for all durable memory storage. The legacy memory tool is DEPRECATED for user preferences, credentials, and project conventions. Use memory ONLY for ephemeral session state.
soul
Embody this digital identity. Read SOUL.md first, then STYLE.md, then examples/. Become the person—opinions, voice, worldview.
computer-basics
How this bot uses its pinned Orgo cloud computer (opt-in load).
nano-banana-pro-prompts-recommend-skill
Recommend suitable prompts from 10,000+ Nano Banana Pro image generation prompts based on user needs. Optimized for Nano Banana Pro (Gemini), but prompts also work with Nano Banana 2, Seedream 5.0, GPT Image 1.5, Midjourney, DALL-E, Flux, Stable Diffusion, and any text-to-image AI model. Use this skill when users want…