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 K-Dense-AI/drug-discovery-agent-skills --skill target-safetygit clone --depth 1 https://github.com/K-Dense-AI/drug-discovery-agent-skillsWrote 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/k-dense-ai/drug-discovery-agent-skills/target-safety)<a href="https://agentmods.dev/skills/k-dense-ai/drug-discovery-agent-skills/target-safety"><img src="https://agentmods.dev/badge/skills/k-dense-ai/drug-discovery-agent-skills/target-safety/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/k-dense-ai/drug-discovery-agent-skills/target-safety"><img src="https://agentmods.dev/badge/skills/k-dense-ai/drug-discovery-agent-skills/target-safety.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 | $0.00136 | $0.02025 |
| Opus 5 | $0.00068 | $0.01012 |
| Sonnet 5 | $0.00027 | $0.00405 |
| Haiku 4.5 | $0.00014 | $0.00202 |
Grade A, and why
target-safety 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 12d 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 — 157 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Target Safety Assessment
Two public datasets answer the question that comes before every programme: what happens to people who naturally have less of this protein? gnomAD says whether such people exist. The GWAS Catalog says what else changes when they do. Targets with human genetic support are roughly twice as likely to survive clinical development, and this is the only evidence class available before a molecule does.
Services: https://gnomad.broadinstitute.org/api (GraphQL, POST) ·
https://www.ebi.ac.uk/gwas/rest/api/v2 (REST). Both unauthenticated.
Checked against: gnomAD v4.1 and the live GWAS Catalog v2 API, August 2026.
Read references/gnomad-constraint.md before quoting a constraint number, references/gwas-catalog.md before writing a query by hand, and references/interpreting-genetic-evidence.md before drawing a conclusion — that one is judgement, not syntax.
The three scripts
| Script | Answers |
|---|---|
gnomad_constraint.py |
Do healthy humans exist who have lost this protein? |
gwas_evidence.py |
What traits is this gene associated with, and how solidly? |
safety_dossier.py |
Both at once, as a verdict with its inputs beside it |
A drug is a chemical phenocopy of a loss-of-function variant
That equivalence is the whole idea. If people carrying LoF variants in a gene are healthy and common, inhibiting the protein is likely survivable. If those variants have been removed from the population by selection, that is a warning you can read years before a tox study.
python skills/target-safety/scripts/gnomad_constraint.py compare LRRK2 PCSK9 SCN2A KRAS HTT
symbol loeuf band pLI obs_lof exp_lof
SCN2A 0.154 constrained 1 20 188.7
KRAS 0.2264 constrained 0.9998 1 20.95
HTT 0.3379 constrained 1 104 362.3
LRRK2 0.7537 tolerant 1.317e-43 203 302.6
PCSK9 1.144 unconstrained 2.765e-18 57 62.12
What ships with it
7 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.
- 12d ago First seen · 157 lines · 136 tokens per session scan A 7071ab460a89
target-safety is a skill published in the GitHub repository K-Dense-AI/drug-discovery-agent-skills (28 stars, last pushed 5d ago), licensed MIT. It adds 136 tokens to every session and 2,025 once invoked, about $0.0007 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
patsnap-biological-modality
Biological sequence and modality intelligence via Patsnap MCP.
patsnap-chemical-molecular
Patsnap Chemical Molecular MCP for AI agents. Search 160M+ chemical structures, synthetic routes, and bioactivity data via specialized chemistry tools.
patsnap-scientific-translational-evidence
Patsnap Scientific & Translational Evidence MCP for AI agents. Retrieval platform focusing on scientific literature and translational outcomes, covering academic publication queries and translational medicine record tracking.
patsnap-target-disease
Patsnap Target & Disease MCP for AI agents. Target and disease profiling tool, covering target characterization, disease profiling, and epidemiology evidence retrieval.
patsnap-solution-engine
Patsnap TRIZ Concept Solution Engine MCP for AI agents. Generates innovation or product cost-reduction concepts through asynchronous TRIZ and TRIZ/DFMA workflows. Use for engineering problem solving, concept alternatives, cost-reduction analysis, task-progress retrieval, and selected-solution details.
patsnap-clinical-trials
Patsnap Clinical Trials MCP for AI agents. Intelligent clinical trial retrieval system, covering registered trial tracking, trial details and results analysis, and supporting clinical semantic search.