Skip to main content

workflow-create

Facilitates the creation and management of reusable skill workflows, enabling users to streamline processes and enhance productivity.

Install this skill

or
0/100

Security score

The workflow-create skill was audited on Jun 15, 2026 and we found 14 security issues across 3 threat categories, including 9 high-severity. Review the findings below before installing.

Categories Tested

Security Issues

high line 121

Template literal with variable interpolation in command context

SourceSKILL.md
1211. If `skill-creator` is already installed (host agent exposes it, or a folder exists under `${SKILLS_ROOT}/skill-creator`), use it. The installed copy is whatever the user last pulled, so upstream up
high line 122

Template literal with variable interpolation in command context

SourceSKILL.md
1222. If it is not present, **retrieve it** from the canonical source (`npx skills add anthropics/skills --skill skill-creator`, or fetch the files into `${SKILLS_ROOT}/skill-creator`), then follow it. T
high line 170

Template literal with variable interpolation in command context

SourceSKILL.md
170- **Claude Code** (the only exception) → `${USER_ROOT}/.claude/skills` (global) or `<project-root>/.claude/skills` (project-local), its native home.
high line 171

Template literal with variable interpolation in command context

SourceSKILL.md
171- **Every other / unknown harness** → `${USER_ROOT}/.agents/skills` (global) or `<project-root>/.agents/skills` (project-local), the standard cross-client home.
high line 173

Template literal with variable interpolation in command context

SourceSKILL.md
173Set `${SKILLS_ROOT}` to the resolved home for the detected harness and chosen scope. This is where the real folders are written — no symlink needed for the host harness to find them. Cross-linking to
high line 184

Template literal with variable interpolation in command context

SourceSKILL.md
184Resolve install locations from the user's home directory or the named project root, not hardcoded machine-specific paths. If the chosen `${SKILLS_ROOT}` does not exist, create it for `create` or `shel
high line 194

Template literal with variable interpolation in command context

SourceSKILL.md
194- **Host is Claude Code** (family written to `.claude/skills`): ask whether they also want these skills available to non–Claude Code agents. Something like: *"Want me to also link these into `~/.agent
high line 195

Template literal with variable interpolation in command context

SourceSKILL.md
195- **Host is another agent** (family written to `.agents/skills`): ask whether they also want these in Claude Code. Something like: *"Want me to also link these into `~/.claude/skills` so Claude Code p
high line 203

Template literal with variable interpolation in command context

SourceSKILL.md
203For public or distributable workflows, write paths with placeholders such as `${USER_ROOT}`, `${SKILLS_ROOT}`, and `${PARENT_SKILL_DIR}`. Avoid absolute paths that include the creator's machine-specif
medium line 295

Template literal with variable interpolation in command context

SourceSKILL.md
295```text
medium line 168

Access to hidden dotfiles in home directory

SourceSKILL.md
1682. **Confirm from the environment (optional sanity check).** Claude Code exports `CLAUDECODE=1` (plus `CLAUDE_CODE_*` vars) into the shell. If you want to corroborate, a quick `env | grep -i claudecod
medium line 194

Access to hidden dotfiles in home directory

SourceSKILL.md
194- **Host is Claude Code** (family written to `.claude/skills`): ask whether they also want these skills available to non–Claude Code agents. Something like: *"Want me to also link these into `~/.agent
medium line 195

Access to hidden dotfiles in home directory

SourceSKILL.md
195- **Host is another agent** (family written to `.agents/skills`): ask whether they also want these in Claude Code. Something like: *"Want me to also link these into `~/.claude/skills` so Claude Code p
low line 105

External URL reference

SourceSKILL.md
105Every generated name must conform to the Agent Skills open standard (https://agentskills.io/specification):
Scanned on Jun 15, 2026
View Security Dashboard
Installation guide →