Skip to content

K-Dense-AI/scientific-agent-skills / skills/clinical-decision-support

clinical-decision-support

Verified

Prepares and validates research-only clinical decision-support evaluation, evidence-profile, cohort, survival, biomarker/model, privacy, and governance artifacts. Supports aggregate or synthetic research documentation and traceability, excluding patient care and live clinical operation.

references/security_validation.md

View on GitHub
# Security Validation Record

Validation date: **2026-07-23** (local project date).

## Baseline

The repository `SECURITY.md` section recorded 11 findings:

- 3 CRITICAL findings for cross-file environment-variable/network exfiltration behavior;
- 1 HIGH finding for transmitting a credential to an external service;
- 5 MEDIUM findings for command chaining, credential flow, undeclared network use, and environment harvesting;
- 2 LOW findings for mandatory cross-skill invocation and dependency concerns.

The affected external schematic wrappers were the now-deleted files
`generate_schematic.py` and `generate_schematic_ai.py`.

## Remediation

- Deleted both schematic-generation scripts.
- Removed all external service, LLM, image, credential, and environment-variable behavior.
- Removed mandatory figures and cross-skill calls.
- Replaced unsafe person-level classification and care-pathway helpers with aggregate or planning-only tools.
- Added bounded local JSON handling, person-level-key rejection, output limits, and deterministic schemas.
- Added static AST tests that reject network libraries, dynamic code execution, credential access, and executable serialization.

## Historical July 2026 Scan Results

- Direct behavioral scan: **SAFE, 0 findings**.
- Pull-request gate with `--fail-on HIGH`: **passed** with 0 CRITICAL and
  0 HIGH. Repeated LLM-assisted runs returned 2–3 LOW findings because the
  analyzer is nondeterministic.

## Accepted LOW Findings

Across repeated runs, the nondeterministic analyzer reported different subsets of
these LOW observations:

1. **Optional `allowed-tools` field absent** — accepted as informational. The
   Agent Skills specification does not require it; compatibility and runtime
   instructions explicitly prohibit network and credential access.
2. **Broad description** — accepted. The scope intentionally covers the related research-evaluation artifacts requested for this safety refresh, while the frontmatter and first body section explicitly exclude care and live operation.
3. **Person-level key filtering uses a denylist** — accepted as a documented limitation, not a privacy guarantee. Input is contractually restricted to synthetic or aggregate schemas, common person-level fields are rejected as defense in depth, and the documentation repeatedly requires qualified privacy review. The filter cannot detect every identifier name or sensitive value and is never presented as de-identification.
4. **Occasional missing-file report** — accepted as an analyzer false
   positive. Some runs invented a `templates/` directory and nonexistent
   `assets/*.md` references. The deterministic
   `test_documented_local_paths_exist` check resolves every documented local
   path and passes.

None of the accepted findings permits network access, sensitive-data handling, or clinical action.

## September 2026 Maintenance

The 2.4 refresh keeps all runtime helpers offline. Invalid aggregate/matrix inputs now withhold results or exports; denominator checks run before suppression; undefined balanced accuracy stays null; CSV text is escaped and dangling output symlinks are rejected. Synthetic regression tests cover these cases and the updated GRADE imprecision label. The July scanner results above are historical, not a claim about a new scan of version 2.4.

## Reproduction

```bash
uv run skill-scanner scan skills/clinical-decision-support --use-behavioral

uv run python scan_pr_skills.py \
  --fail-on HIGH \
  --output /tmp/clinical-decision-support-pr-scan.md \
  skills/clinical-decision-support
```
rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.Stars: 198kUpdated Claude CodeCodexVerifiedInstall
claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file).Stars: 198kUpdated Claude CodeCodexUnverifiedInstall
doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.Stars: 198kUpdated Claude CodeCodexVerifiedInstall