Claude Code, Skill

Galact Skills

一套自用的 Agent Skill（SKILL.md 标准，跨 Claude Code / Codex 等）集合。

Compatibility: Not stated.

Grade D

Two or more Medium findings, or a High. Only try it in a sandbox.

Graded by PlugVet23/10/2026Rubric v2

Not a guarantee. A hygiene check of the source at the commit shown, not a security audit.

NETWORK_HOST_FROM_INPUT SECRETS_PATH_CREDENTIAL

Why this grade

NETWORK_HOST_FROM_INPUT High
A file takes a URL or endpoint as input and has no fixed host list.
h1_data/pipeline/probe.py, line 54 with urllib.request.urlopen(req, timeout=300) as resp:
SECRETS_PATH_CREDENTIAL Medium
Mentions a credential store path (~/.ssh, id_rsa, .npmrc, .aws/credentials, credentials.json, .netrc).
skills/hunt-command-injection/reference.md, line 20 - **git archive/checkout 注入 `--output=`** 覆盖任意文件（如 `~/.ssh/authorized_keys`）→ SSH 登录 → RCE：#587854、#658013。

Any High, or two or more Medium findings, give D. One Medium gives C. Low findings or a security Note give B. Each code names a finding behind the letter.

If you think this grade is wrong, email [email protected] with the entry link and your evidence. We aim to acknowledge within two working days and decide within seven. If we confirm an error, we aim to correct it within 24 hours of confirming it. These are aims, not guarantees.

Grades are a hygiene check of the source at the commit shown, on the date checked, not a security audit or a guarantee. The free directory grades are automated. Before release, a PlugVet engineer reviews every paid report and any grade we hold back, including all F grades, against the source code and the evidence we quote.

What was checked

Plain notes from the hygiene checks.

Permissions
Concern
can delete files.
PERMISSION_DELETE_FILES
h1_data/pipeline/build_batches.py, line 36
os.remove(os.path.join(args.out, f))
h1_data/pipeline/run_round2.py, line 177
os.remove(f)
skills/hunt-auth-bypass/tools/common.py, line 63
p = subprocess.run(cmd, capture_output=True, text=True, timeout=max_time + 3)
Show more
skills/hunt-auth-bypass/tools/validate.sh, line 55
p = subprocess.run(["curl", "-sS", "-i", "--max-time", "8", purl],
skills/hunt-cache-poisoning/tools/common.py, line 63
p = subprocess.run(cmd, capture_output=True, text=True, timeout=max_time + 3)
skills/hunt-cache-poisoning/tools/hunt_cache.py, line 32
p = subprocess.run(cmd, capture_output=True, text=True, timeout=max_time + 3)
skills/hunt-cache-poisoning/tools/validate.sh, line 55
p = subprocess.run(["curl", "-sS", "-i", "--max-time", "8", purl],
skills/hunt-command-injection/tools/common.py, line 63
p = subprocess.run(cmd, capture_output=True, text=True, timeout=max_time + 3)
skills/hunt-command-injection/tools/validate.sh, line 55
p = subprocess.run(["curl", "-sS", "-i", "--max-time", "8", purl],
skills/hunt-csrf/tools/common.py, line 63
p = subprocess.run(cmd, capture_output=True, text=True, timeout=max_time + 3)
Install and update
Pass
No install-time script found, and dependencies are pinned or absent.
Network
Serious
accepts a host from input, with no fixed list in this file.
NETWORK_HOST_FROM_INPUT
h1_data/pipeline/probe.py, line 54
with urllib.request.urlopen(req, timeout=300) as resp:
h1_data/pipeline/route_singletons.py, line 74
with urllib.request.urlopen(req, timeout=timeout) as r:
h1_data/pipeline/run_round2.py, line 94
with urllib.request.urlopen(req, timeout=timeout) as resp:
Show more
h1_data/pipeline/run_scoring.py, line 85
with urllib.request.urlopen(req, timeout=timeout) as resp:
skills/hunt-auth-bypass/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-auth-bypass/tools/recon.sh, line 28
curl -sS --max-time 15 -L "$TARGET" -o "$PAGE" || echo "[recon] 目标抓取失败，继续" >&2
skills/hunt-cache-poisoning/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-cache-poisoning/tools/recon.sh, line 28
curl -sS --max-time 15 -L "$TARGET" -o "$PAGE" || echo "[recon] 目标抓取失败，继续" >&2
skills/hunt-command-injection/reference.md, line 12
;nslookup <OAST>; | curl http://<OAST>/hit # 带外确认
skills/hunt-command-injection/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-command-injection/tools/hunt_cmdi.py, line 6
带外判活：注 nslookup/curl <OAST>，命中需你在 OAST 侧核对。
skills/hunt-command-injection/tools/hunt_cmdi.py, line 32
("oast-curl", ";curl http://%s/hit;" % oast, "oast"),
skills/hunt-command-injection/tools/recon.sh, line 29
curl -sS --max-time 15 -L "$TARGET" -o "$PAGE" || echo "[recon] 目标抓取失败，继续" >&2
skills/hunt-csrf/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-csrf/tools/recon.sh, line 28
curl -sS --max-time 15 -L "$TARGET" -o "$PAGE" || echo "[recon] 目标抓取失败，继续" >&2
skills/hunt-deserialization/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-deserialization/tools/recon.sh, line 28
curl -sS --max-time 15 -L "$TARGET" -o "$PAGE" || echo "[recon] 目标抓取失败，继续" >&2
skills/hunt-nodejs-permission-bypass/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-open-redirect/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-open-redirect/tools/recon.sh, line 28
curl -sS --max-time 15 -L "$TARGET" -o "$PAGE" || echo "[recon] 目标抓取失败，继续" >&2
skills/hunt-path-traversal/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-path-traversal/tools/hunt_pathtrav.py, line 54
"""path traversal 需要看 body，用 curl 直接取 body（http_probe 丢弃 body）。"""
skills/hunt-path-traversal/tools/recon.sh, line 29
curl -sS --max-time 15 -L "$TARGET" -o "$PAGE" || echo "[recon] 目标抓取失败，继续" >&2
skills/hunt-prototype-pollution/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-prototype-pollution/tools/recon.sh, line 28
curl -sS --max-time 15 -L "$TARGET" -o "$PAGE" || echo "[recon] 目标抓取失败，继续" >&2
skills/hunt-request-smuggling/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-request-smuggling/tools/hunt_smuggling.py, line 53
s = socket.create_connection((host, port), timeout=timeout)
skills/hunt-request-smuggling/tools/recon.sh, line 33
curl -sS -D "$HDR" -o /dev/null --max-time 15 "$TARGET" || echo "[recon] 抓取失败，继续" >&2
skills/hunt-sqli/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-sqli/tools/recon.sh, line 28
curl -sS --max-time 15 -L "$TARGET" -o "$PAGE" || echo "[recon] 目标抓取失败，继续" >&2
skills/hunt-ssrf/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-ssrf/tools/recon.sh, line 44
curl -sS --max-time 15 -L "$TARGET" -o "$PAGE" || echo "[recon] 目标抓取失败（可能需登录/非 GET），继续用空页面" >&2
skills/hunt-xss/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-xss/tools/hunt_xss.py, line 33
def fetch(url, max_time=8):
skills/hunt-xss/tools/hunt_xss.py, line 82
body = fetch(url)
skills/hunt-xss/tools/recon.sh, line 28
curl -sS --max-time 15 -L "$TARGET" -o "$PAGE" || echo "[recon] 目标抓取失败，继续" >&2
skills/hunt-xxe/tools/common.py, line 46
"""用 curl 发一次探测，返回 {status, size, time, headers, ok, cmd, stderr}。
skills/hunt-xxe/tools/recon.sh, line 28
curl -sS --max-time 15 -L "$TARGET" -o "$PAGE" || echo "[recon] 目标抓取失败，继续" >&2
Secrets
Concern
reads a credential or environment file.
SECRETS_ENV_READ SECRETS_PATH_CREDENTIAL SECRETS_PATH_DOTENV
h1_data/pipeline/probe.py, line 19
ap.add_argument("--api-key", default=os.environ.get("OPENAI_API_KEY", ""))
h1_data/pipeline/run_round2.py, line 268
ap.add_argument("--api-key", default=os.environ.get("OPENAI_API_KEY", ""))
skills/hunt-command-injection/reference.md, line 20
- **git archive/checkout 注入 `--output=`** 覆盖任意文件（如 `~/.ssh/authorized_keys`）→ SSH 登录 → RCE：#587854、#658013。
Show more
h1_data/pipeline/run_scoring.py, line 142
ap.add_argument("--api-key", default=os.environ.get("OPENAI_API_KEY", ""))
skills/recognize-attack-surface/SKILL.md, line 56
- 服务：调试端点、管理面暴露、默认账号、目录列举、`.git`/`.env`/备份可访问。
Agent instructions
Pass
No instruction to ignore earlier instructions or to hide a step.
Obfuscation
Pass
No encoded or minified payload found.
Licence
Pass
A licence file is present (LICENSE).
Maintainer activity
Pass
The latest commit is within the last 365 days.

Get the full report on this tool, £49

Note

Hygiene check of commit 100995116742. can delete files. accepts a host from input, with no fixed list in this file. reads a credential or environment file.

Date and source

Date checked: 23 Oct 2026

View source

github.com

Commit graded: 1009951167421bad5a4342104eff962f6e1f8903

This link goes to the original host. PlugVet does not host or check what you download from there, and installing it is your own decision and risk.

Read how this letter is chosen

What was checked

Permissions

Install and update

Network

Secrets

Agent instructions

Obfuscation

Licence

Maintainer activity

Note

Date and source