스킬 gardener

📦

gardener

Name: gardener
Author: drix10

콘텐츠 리비전 r1 안전 ⚙️ 외부 명령어

Maintain Accurate Agent Context

Stale context files mislead agents and increase maintenance risk. Gardener verifies documentation against code, repairs references, and keeps risk baselines controlled.

지원: Claude Codex Code(CC)

🥉 77 브론즈

Maintain this Skill? Claim it 🥉 77 브론즈

내 에이전트로 설치

이 요청을 에이전트에 복사하세요. 공식 Skill 페이지와 매니페스트가 포함되어 있습니다.

에이전트 요청

Review the Skillstore skill "gardener" from https://skillstore.io/skills/drix10-gardener.md and its manifest at https://skillstore.io/api/skills/drix10-gardener/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.

에이전트는 계속해서 계획을 표시하고 보안 정책에서 요구하는 확인을 요청해야 합니다.

에이전트가 읽기 쉬운 리소스

AI 에이전트, 크롤러, 스크립트가 전체 페이지 대신 정리된 컨텍스트만 필요할 때 이 링크를 사용하세요.

Markdown 상세 GET /skills/drix10-gardener.md 서명된 manifest GET /api/skills/drix10-gardener/manifest 서명된 lockfile GET /api/skills/drix10-gardener/lockfile

테스트해 보기

"gardener" 사용 중입니다. Check whether our agent context is stale after the latest merge.

예상 결과:

Context health: two stale references found in AGENTS.md. No files changed. Next action: run the documentation repair workflow.

"gardener" 사용 중입니다. Review new risk surfaces without accepting anything.

예상 결과:

Protected path candidate: deployment credentials directory.
Risk boundary candidate: release automation scripts.
Human decision required before any baseline update.

"gardener" 사용 중입니다. Repair the flagged documentation and verify it again.

예상 결과:

Updated verified claims and manifest references. The final context check passed, with one implementation issue opened for a missing lint rule.

보안 감사

안전

v1 • 2026. 9. 27. 버전별 보고서 열기

Most static findings are false positives caused by Markdown backticks around paths, identifiers, and tool names. Two findings are confirmed because the skill directs agents to run an unpinned npx package and a git command with a repository-derived path.

스캔된 파일

분석된 줄 수

검토 항목

오탐 무시됨

기능 검토 항목 (2)

이 스킬에 대해 예상될 수 있는 실제 로컬 기능이므로 검토가 필요하지만, 확인된 악의적 동작으로는 집계되지 않습니다.

중간

SKILL.md:18

Ruby/shell backtick execution

Outside Pi, use `npx agent-flow doctor`, `npx agent-flow audit-risk` and `npx agent-flow baseline ac

The line explicitly directs the agent to execute unpinned npx agent-flow commands. npx may download and run registry code, creating a supply-chain and external-command risk.

중간

SKILL.md:34

Ruby/shell backtick execution

1. List the docs: `docs/`, `design/`, `adr/`, `rfcs/`, root `*.md`. Get each one's last commit date

The line directly instructs the agent to run git log for repository-derived paths. The option terminator reduces option injection, but unsafe shell interpolation of a path could still execute metacharacters.

감사자: codex

이 보고서 공유 및 인용

버전이 지정된 평가 보고서, 중립 배지, 임베드 카드 및 인용을 공유하세요. Skillstore는 근거를 보고하지만 이 Skill의 안전 여부를 판단하지 않습니다.

버전별 보고서 열기

보고서 링크 복사

https://skillstore.io/skills/drix10-gardener/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown 배지

[![Skillstore security assessment](https://skillstore.io/badges/skills/drix10-gardener/security.svg)](https://skillstore.io/skills/drix10-gardener?utm_source=security_passport_badge)

HTML 배지

<a href="https://skillstore.io/skills/drix10-gardener?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/drix10-gardener/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

임베드 카드

<iframe src="https://skillstore.io/embed/skills/drix10-gardener.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

학술 인용 (APA · BibTeX · CFF)

APA 인용

drix10. (2026). gardener security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/drix10-gardener/audits/1

BibTeX 인용

@techreport{drix10-drix10-gardener-2026,
  author = {drix10},
  title = {gardener security audit report (audit version 1)},
  institution = {Skillstore},
  year = {2026},
  number = {1},
  url = {https://skillstore.io/skills/drix10-gardener/audits/1},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "gardener security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
  - name: "drix10"
date-released: "2026-09-27"
url: "https://skillstore.io/skills/drix10-gardener/audits/1"
identifiers:
  - type: other
    value: "skillstore:drix10-gardener:audit:1"
    description: "Skillstore immutable audit report identifier"

Skillstore 점수

이 점수의 이유 증거 신뢰도: 중간

아키텍처

유지보수성

콘텐츠

커뮤니티

사양 준수

만들 수 있는 것

Check Context Health

Inspect context files for stale claims, missing references, invalid timestamps, and unresolved placeholders.

Review Risk Baselines

Evaluate new risk surfaces and prepare focused choices for human approval.

Refresh Documentation Indexes

Classify documentation as current, stale, missing, or archived using repository history.

이 프롬프트를 사용해 보세요

Run a Context Check

Run the context health check. Report every failed check and provide the exact next command without changing files.

Synchronize Documentation

Update the documentation index from current repository files and commit dates. Show stale, missing, and archived entries before applying edits.

Repair Stale Context

Find stale context claims, verify each against current code, repair supported statements, update references, and rerun validation.

Complete Repository Gardening

Run synchronization, risk audit, diagnosis, and repair in order. Summarize fixes, required human decisions, and proposed enforcement mechanisms.

모범 사례

Review the code behind every flagged claim before changing documentation or timestamps.
Keep human approval for risk acceptance and protected write operations.
Run the final health check after every repair and explain any remaining failure.

피하기

Do not refresh verification timestamps before validating claims against current code.
Do not accept secrets into a risk baseline or treat heuristic matches as final verdicts.
Do not delete superseded documentation, edit protected paths, or modify source code through this role.

자주 묻는 질문

What files can Gardener edit?

Under the guarded role, it can edit Markdown files and CONTEXT_MANIFEST.json. Protected paths remain unavailable.

Does Gardener change application code?

No. It opens an implementation issue when a fix requires source code, configuration, hooks, or refactoring.

Can Gardener repair documentation automatically?

It can edit verified documentation, but stale repair certification and other protected operations require human confirmation.

How does the risk audit work?

It treats detected patterns as leads, groups new surfaces, and asks a human to choose the appropriate response.

What happens to outdated documents?

Gardener marks superseded documents as archived. It does not delete documentation.

Can it run outside the Pi environment?

Yes. The skill provides agent-flow command alternatives, but the required package should be installed locally or pinned to a trusted version.

개발자 세부 정보

작성자

drix10

라이선스

MIT

Skillstore 리비전

버전 안내

작성자가 버전을 선언하지 않았습니다.

리포지토리

https://github.com/drix10/agent-flow/tree/aac4c5c023b25d073997626804c355ed11348feb/skills/gardener

참조

8b5ad1599e7c55e9b6c627bca3f556f6e9061cb1

유지보수 최신성

2026. 9. 30.

사용량

0 다운로드 · 0 조회

파일 구조

📄 SKILL.md

gardener

내 에이전트로 설치

에이전트가 읽기 쉬운 리소스

테스트해 보기

보안 감사

위험 요인

보고서 링크 복사

Markdown 배지

HTML 배지

임베드 카드

APA 인용

BibTeX 인용

CITATION.cff

Skillstore 점수

만들 수 있는 것

Check Context Health

Review Risk Baselines

Refresh Documentation Indexes

이 프롬프트를 사용해 보세요

모범 사례

피하기

자주 묻는 질문

개발자 세부 정보