← All skills
reproducer
57 / 100 · Review before installing
Check before installing. Some of what it can do is expected: a security skill usually runs scanners and checks websites.
By apache/magpie (GitHub) · security · 146 stars · scanned at commit 8d24b52
For a single `<issue-tracker>` issue identifying a code-level
bug, extract the reporter's example code from the issue body,
adapt it to run on the current `<default-branch>`, execute via
`<runtime>`, and compose a `verdict.json` describing the
observed behaviour vs the expected failure. Read-only on the
tracker — produces evidence, never posts. Invoked by
`issue-triage` and `issue-reassess`; can also be run standalone.
Works with any agent that supports SKILL.md skills.
What this skill can do
Runs commands on your computer
The author does not mention this. Expected for a security skill.
Found in 1 place.
Downloads and installs other software
The author does not mention this.
Found in 32 places.
Show 17 more
- docs/setup/secure-agent-setup.md, line 217
sudo apt-get install --no-install-recommends \ - docs/setup/secure-agent-setup.md, line 219
npm install -g --no-save @anthropic-ai/claude-code@latest - docs/setup/secure-agent-setup.md, line 321
sudo apt-get install --no-install-recommends \ - docs/setup/secure-agent-setup.md, line 342
sudo dnf install \ - docs/setup/secure-agent-setup.md, line 348
optional. If you want socat, `brew install socat` (current Homebrew - docs/setup/secure-agent-setup.md, line 356
npm install -g --no-save @anthropic-ai/claude-code@latest - docs/setup/secure-agent-setup.md, line 378
sudo apt-get install --no-install-recommends \ - docs/setup/secure-agent-setup.md, line 532
"uvx --from ~/.claude/magpie/vetted-ops vetted-op-read *", - docs/setup/secure-agent-setup.md, line 534
"uvx --from ~/.claude/magpie/vetted-ops vetted-op-tracker *", - docs/setup/secure-agent-setup.md, line 536
"uvx --from ~/.claude/magpie/vetted-ops vetted-op *", - docs/setup/secure-agent-setup.md, line 537
"uvx --from ~/.claude/magpie/adversarial-review adversarial-review *" - docs/setup/secure-agent-setup.md, line 697
"Bash(uvx --from ~/.claude/magpie/vetted-ops vetted-op-read *)", // same dispatcher, the form bulk gatherer agents use - docs/setup/secure-agent-setup.md, line 733
"Bash(pip install --upgrade *)", "Bash(uv self update *)", - docs/setup/secure-agent-setup.md, line 762
"Bash(uvx --from ~/.claude/magpie/vetted-ops vetted-op *)", - docs/setup/secure-agent-setup.md, line 764
"Bash(uvx --from ~/.claude/magpie/vetted-ops vetted-op-tracker *)", - docs/setup/secure-agent-setup.md, line 1004
1. `permissions.allow` carries `vetted-op-read` for both launchers (`uvx --from` and `uv run --project`) with the literal `~/.claude/magpie/vetted-ops` base. - docs/setup/secure-agent-setup.md, line 1005
2. No `permissions.ask` rule, in any scope, matches a `vetted-op-read` command in either launcher form; a broad rule such as `Bash(uvx *)` counts.
and 12 more places
Reads your saved credentials, such as SSH keys or .env files
This is disclosed by the author in the skill's SKILL.md or README.
Found in 15 places.
Show 12 more
- docs/setup/secure-agent-setup.md, line 721
"Read(~/.docker/**)", "Read(~/.kube/**)", - docs/setup/secure-agent-setup.md, line 725
"Read(~/.config/gcloud/**)", "Read(~/.azure/**)", - docs/setup/secure-agent-setup.md, line 1824
echo '{"tool_name":"Bash","tool_input":{"command":"ls ~/.aws","description":"check aws creds","dangerouslyDisableSandbox":true}}' \ - docs/setup/secure-agent-setup.md, line 2646
ssh-add -L > ~/.ssh/id_yubikey.pub - docs/setup/secure-agent-setup.md, line 2648
git config --global user.signingkey ~/.ssh/id_yubikey.pub - docs/setup/secure-agent-setup.md, line 3800
test -r ~/.aws/credentials && echo readable # → nothing (denied or absent; see the 7a sentinel rule) - docs/setup/secure-agent-setup.md, line 3871
`~/.aws/credentials`, the `~/.ssh/id_*` private keys (list the readable ones with - docs/setup/secure-agent-setup.md, line 3872
`find -H ~/.ssh -maxdepth 1 -name 'id_*' ! -name '*.pub' -exec test -r {} \; -print`, - docs/setup/secure-agent-setup.md, line 3874
`~/.ssh`; its error separates absent from denied except under - docs/setup/secure-agent-setup.md, line 3876
`~/.config/gh/hosts.yml`, `~/.netrc`, `~/.docker/config.json`: - docs/setup/secure-agent-setup.md, line 4332
`dangerouslyDisableSandbox: true` to run `ls ~/.aws/`"*. The - docs/setup/secure-agent-setup.md, line 4355
![Sandboxed Bash call running `python3 -c 'os.listdir(os.path.expanduser("~/.aws/"))'`; the inner syscall fails with PermissionError: [Errno 1] Operation not pe
Sends data to these websites: example.com
This is disclosed by the author in the skill's SKILL.md or README. Expected for a security skill.
Found in 5 places.
Show 2 more
Changes files outside its own folder
The author does not mention this.
Found in 28 places.
Show 17 more
- docs/setup/secure-agent-setup.md, line 1902
mkdir -p ~/.claude/scripts/guards.d - docs/setup/secure-agent-setup.md, line 1909
cp /path/to/magpie/skills/*/guards/*.py ~/.claude/scripts/guards.d/ 2>/dev/null || true - docs/setup/secure-agent-setup.md, line 2069
mkdir -p ~/.claude/scripts - docs/setup/secure-agent-setup.md, line 2286
mkdir -p ~/.claude/scripts - docs/setup/secure-agent-setup.md, line 2438
mkdir -p ~/.claude/scripts - docs/setup/secure-agent-setup.md, line 2646
ssh-add -L > ~/.ssh/id_yubikey.pub - docs/setup/secure-agent-setup.md, line 2770
mkdir -p ~/.claude/scripts - docs/setup/secure-agent-setup.md, line 2883
ln -s gpg-touch-overlay.sh ~/.claude/scripts/gpg-touch-wrap-ssh-keygen - docs/setup/secure-agent-setup.md, line 2933
mkdir -p ~/.claude/scripts/shims - docs/setup/secure-agent-setup.md, line 2935
ln -sfn ~/.claude/scripts/gpg-touch-overlay.sh ~/.claude/scripts/shims/"$p" - docs/setup/secure-agent-setup.md, line 3010
mkdir -p ~/.claude/scripts - docs/setup/secure-agent-setup.md, line 3013
rm -f ~/.claude/scripts/container-gateway-hook.sh - docs/setup/secure-agent-setup.md, line 3014
cp /path/to/magpie/tools/agent-isolation/container-gateway-hook.sh ~/.claude/scripts/ - docs/setup/secure-agent-setup.md, line 3016
mkdir -p ~/.claude/scripts/container-gateway/src - docs/setup/secure-agent-setup.md, line 3224
podman run --rm -v "$HOME/.ssh:/x" alpine true # expected: 403 container-gateway: bind-mount … - docs/setup/secure-agent-setup.md, line 3309
mkdir -p ~/.claude - docs/setup/secure-agent-setup.md, line 3311
mv ~/.claude/CLAUDE.md ~/.claude/CLAUDE.md.bak
and 8 more places
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'reproducer' ] || [ -L ~/.agents/skills/'reproducer' ]; then echo 'A skill folder named reproducer already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/apache/magpie' "$d" && git -C "$d" checkout --quiet '8d24b52457f7a2731d7cc3329af96430a16715fd' -- 'plugins/magpie-issue/skills/reproducer' && mkdir -p ~/.agents/skills && cp -R "$d"/'plugins/magpie-issue/skills/reproducer' ~/.agents/skills/'reproducer' && rm -rf "$d"; fi
Codex reads ~/.agents/skills directly. For Claude Code, Cursor, Copilot, Gemini CLI, Antigravity or OpenCode, also link it. For Claude Code: ln -s ~/.agents/skills/'reproducer' ~/.claude/skills/'reproducer' (other agents: ~/.cursor/skills, ~/.copilot/skills, ~/.gemini/skills, ~/.gemini/config/skills, ~/.config/opencode/skills). Or install through the finder skill, which does both.
Badge for your README
The badge for apache/magpie shows the score of its lowest-scoring skill.
[](https://zalavijaysinh.github.io/skilltrust/repos/apache/magpie/)
Score history
| Date | Version | Score |
|---|
| 2026-11-13 | 8d24b52 | 57 |
| 2026-11-09 | 3fd660b | 57 |
| 2026-11-07 | c099f3f | 85 |