interface

From the files and frontmatter at commit 7284097. This is how much a skill can do, not whether it is risky.

• Downloads software

  It fetches programs or packages from the internet, or from a package source other than the default.

  skill.md:77 (opens in a new tab) · api.md:934 (opens in a new tab) · landing-sandbox.md:141 (opens in a new tab)

• Installs more skills

  It tells the agent to install other skills, plugins or tool servers.

  skill.md:139 (opens in a new tab)

• Runs shell commands

  Its instructions or files give the agent commands to run on your computer.

  skill.md:77 (opens in a new tab) · api.md:934 (opens in a new tab) · landing-sandbox.md:141 (opens in a new tab) and 1 more

• Reaches the network

  It sends or fetches data over the internet, for example with curl, an HTTP library or a package install.

  skill.md:77 (opens in a new tab) · api.md:934 (opens in a new tab) · landing-sandbox.md:141 (opens in a new tab)

• Names web addresses

  It mentions outside web addresses the agent may visit. We list them but don't follow them.

  Outside links we did not follow

Pattern check: 3 items to review · commit 7284097 · 1 Nov 2026

• R01 · HighDownload piped into a shell

  skill.md:77 (opens in a new tab)

  `~/.claude/skills/treg/SKILL.md` right after installing the CLI — so `curl {BASE}/install.sh | sh`

• R01 · HighDownload piped into a shell

  api.md:934 (opens in a new tab)

  and `curl {BASE}/install.sh | sh` runs without `-L`. The legacy names also stay in MCP's

• R01 · HighDownload piped into a shell

  landing-sandbox.md:141 (opens in a new tab)

  `curl … /install.sh | sh` now works for anyone with no repo/git access needed. The **Getting started**

We checked every text file in this skill, at commit 7284097 on 1 Nov 2026, against 45 published patterns such as download-and-run commands, hidden text and credential access. We don't run the skill, we don't read the web pages it points to, and we can't judge everything its instructions might lead an agent to do. In published tests, pattern checks like this one missed most malicious skills. Read the skill before you install it — especially anything it asks you to download or run.

How the check worksReport an incorrect finding

Each agent reads SKILL.md a little differently. We check the file against each one's published rules; we don't run the skill in any agent.

• Agent Skills specFails

  The open format from agentskills.io, held to what the strictest agents enforce, so one folder loads everywhere.

  • Error

    The file is named skill.md. Codex and GitHub Copilot need SKILL.md exactly, though the spec's validator accepts lowercase. F1

    Fix: Rename the file to SKILL.md.

  • Error

    There is no name field. The spec requires one. N1

    Fix: Add name: interface.

  • Error

    There is no description. The spec requires one, and agents pick skills by it. D1

    Fix: Add description: with what the skill does and when to use it.

  • Warning

    Fields outside the spec: title, status, sources and related. Uploads to the Claude apps and API, and the spec's validator, reject the file; most agents ignore these fields. O1

    Fix: Delete them for a portable skill, or keep them if you only target the agents that read them.

  • Warning

    Paths that may not work once the skill is installed: line 102: ../../MINIMAX-PLUGIN.md (outside the skill folder), line 136: ../../DSH-PLUGIN.md (outside the skill folder), line 140: ../../CLAUDE-PLUGIN.md (outside the skill folder) and line 157: ../architecture/feedback.md (outside the skill folder). B6

    Fix: Use forward-slash paths relative to SKILL.md, and copy shared files into the skill.

  • Note

    There is no license field, so readers can't tell whether they may reuse the skill. O2

    Fix: Add license: with an SPDX name such as MIT or Apache-2.0, or the name of a bundled licence file.

  • Note

    Reference files that SKILL.md doesn't link: catalog-review-proposal.md, dashboard.md (linked only from api.md), enrich-arena.md (linked only from seo.md), env-import.md (linked only from api.md), landing-sandbox.md (linked only from api.md) and 4 more. Anthropic suggests linking every reference file directly from SKILL.md. B4

    Fix: Link each one from SKILL.md and say when to read it.

  • Note

    Reference files over 100 lines with no table of contents at the top: api.md (1149 lines), cli.md (766 lines), dashboard.md (245 lines), enrich-arena.md (845 lines), env-import.md (258 lines) and 6 more. B5

    Fix: Add a short list of the sections at the top of each.

• Claude CodePasses with warnings

  Claude Code reads its own extra fields, ignores unknown ones, and still loads a skill whose frontmatter is broken, without its fields.

  • Warning

    The file is named skill.md. Claude Code opens SKILL.md by name, so it finds this file only on disks that ignore case, such as the macOS and Windows defaults. F1

    Fix: Rename the file to SKILL.md.

  • Warning

    There is no description. Claude Code uses the body's first line instead: "# The `tools-registry` skill". D1

    Fix: Add description: with what the skill does and when to use it.

  • Warning

    Paths that may not work once the skill is installed: line 102: ../../MINIMAX-PLUGIN.md (outside the skill folder), line 136: ../../DSH-PLUGIN.md (outside the skill folder), line 140: ../../CLAUDE-PLUGIN.md (outside the skill folder) and line 157: ../architecture/feedback.md (outside the skill folder). B6

    Fix: Use forward-slash paths relative to SKILL.md, and copy shared files into the skill.

  • Note

    There is no name field. Claude Code uses the folder name, interface. N1

    Fix: Add name: interface.

  • Note

    Claude Code ignores fields it doesn't know: title, status, sources and related. O1

    Fix: Check the spelling against Claude Code's field list, or delete them.

  • Note

    Reference files that SKILL.md doesn't link: catalog-review-proposal.md, dashboard.md (linked only from api.md), enrich-arena.md (linked only from seo.md), env-import.md (linked only from api.md), landing-sandbox.md (linked only from api.md) and 4 more. Anthropic suggests linking every reference file directly from SKILL.md. B4

    Fix: Link each one from SKILL.md and say when to read it.

  • Note

    Reference files over 100 lines with no table of contents at the top: api.md (1149 lines), cli.md (766 lines), dashboard.md (245 lines), enrich-arena.md (845 lines), env-import.md (258 lines) and 6 more. B5

    Fix: Add a short list of the sections at the top of each.

• Claude apps & API uploadFails

  Uploading to claude.ai or the Claude API, or packaging with Anthropic's script, accepts only the six spec fields and stops on anything else.

  • Error

    There is no name field. Anthropic's validator stops with "Missing 'name' in frontmatter". N1

    Fix: Add name: interface.

  • Error

    There is no description. Anthropic's validator stops with "Missing 'description' in frontmatter". D1

    Fix: Add description: with what the skill does and when to use it.

  • Error

    Upload and packaging fail on fields outside the spec: title, status, sources and related. O1

    Fix: Delete them, and keep your own data under metadata as text values.

  • Warning

    The file is named skill.md. Anthropic's API docs and packaging script use SKILL.md; the claude.ai Help Center writes skill.md, so the apps may accept it. F1

    Fix: Rename the file to SKILL.md.

  • Warning

    Paths that may not work once the skill is installed: line 102: ../../MINIMAX-PLUGIN.md (outside the skill folder), line 136: ../../DSH-PLUGIN.md (outside the skill folder), line 140: ../../CLAUDE-PLUGIN.md (outside the skill folder) and line 157: ../architecture/feedback.md (outside the skill folder). B6

    Fix: Use forward-slash paths relative to SKILL.md, and copy shared files into the skill.

  • Note

    Reference files that SKILL.md doesn't link: catalog-review-proposal.md, dashboard.md (linked only from api.md), enrich-arena.md (linked only from seo.md), env-import.md (linked only from api.md), landing-sandbox.md (linked only from api.md) and 4 more. Anthropic suggests linking every reference file directly from SKILL.md. B4

    Fix: Link each one from SKILL.md and say when to read it.

  • Note

    Reference files over 100 lines with no table of contents at the top: api.md (1149 lines), cli.md (766 lines), dashboard.md (245 lines), enrich-arena.md (845 lines), env-import.md (258 lines) and 6 more. B5

    Fix: Add a short list of the sections at the top of each.

  • Note

    Line 133 names an MCP tool as mcp__treg__call. For the Claude apps and API, Anthropic's guidance is Server:tool, such as GitHub:create_issue. B7

    Fix: Write it as treg:call, using the server's own name.

• OpenAI (Codex & API)Fails

  Codex finds SKILL.md by its exact name and needs a description; OpenAI's Skills API checks the frontmatter against the spec and caps uploads.

  • Error

    The file is named skill.md. Codex looks only for SKILL.md, though OpenAI's upload matches the name in any case. F1

    Fix: Rename the file to SKILL.md.

  • Error

    There is no name field. Codex falls back to the folder name, but OpenAI's upload follows the spec, which requires one. N1

    Fix: Add name: interface.

  • Error

    There is no description. Codex skips a skill without one. D1

    Fix: Add description: with what the skill does and when to use it.

  • Warning

    Fields outside the spec: title, status, sources and related. Codex ignores them; OpenAI doesn't say whether its upload accepts them. O1

    Fix: Delete them, or move your own data under metadata as text values.

  • Warning

    Paths that may not work once the skill is installed: line 102: ../../MINIMAX-PLUGIN.md (outside the skill folder), line 136: ../../DSH-PLUGIN.md (outside the skill folder), line 140: ../../CLAUDE-PLUGIN.md (outside the skill folder) and line 157: ../architecture/feedback.md (outside the skill folder). B6

    Fix: Use forward-slash paths relative to SKILL.md, and copy shared files into the skill.

• VS Code & CopilotFails

  VS Code reads skills one folder below .github/skills, .claude/skills or .agents/skills and falls back to the folder name when the name is missing or differs; GitHub's Copilot docs require the exact name SKILL.md.

  • Error

    The file is named skill.md. GitHub Copilot requires SKILL.md, and VS Code finds this file only on disks that ignore case. F1

    Fix: Rename the file to SKILL.md.

  • Warning

    There is no name field. VS Code uses the folder name; GitHub's Copilot docs list the name as required. N1

    Fix: Add name: interface.

  • Warning

    There is no description. VS Code still lists the skill, but the agent has nothing to match it against. D1

    Fix: Add description: with what the skill does and when to use it.

  • Warning

    Paths that may not work once the skill is installed: line 102: ../../MINIMAX-PLUGIN.md (outside the skill folder), line 136: ../../DSH-PLUGIN.md (outside the skill folder), line 140: ../../CLAUDE-PLUGIN.md (outside the skill folder) and line 157: ../architecture/feedback.md (outside the skill folder). B6

    Fix: Use forward-slash paths relative to SKILL.md, and copy shared files into the skill.

  • Note

    VS Code ignores fields it doesn't support: title, status, sources and related. O1

    Fix: Delete them unless another agent you target reads them.

14 files