Agent Compatibility
Which AI agents are verified hands-on with speckit-security, and which are inferred from Spec Kit's supported list.
speckit-security is agent-neutral, the extension ships
generic Markdown command files and Spec Kit handles the translation
into each AI agent's native format at install time.
Verified end-to-end
These agents have been smoke-tested with a full install cycle:
specify init --ai <agent>, specify extension add --dev ...,
specify extension list, and execution of the gate-check.sh script
from inside the installed project.
| Agent | Install location | Format |
|---|---|---|
| Claude Code | .claude/skills/speckit-tekimax-security-*/SKILL.md | Markdown skill with frontmatter |
| OpenCode | .opencode/command/speckit.tekimax-security.*.md | Markdown command |
| GitHub Copilot | .github/agents/*.agent.md + .github/prompts/*.prompt.md | Agent + prompt files |
| Gemini CLI | .gemini/commands/speckit.tekimax-security.*.toml | TOML command |
| Cursor | .cursor/skills/speckit-tekimax-security-*/SKILL.md | Markdown skill |
All eight speckit-security commands register cleanly on each
agent, and the gate-check script runs end-to-end after install.
Supported by Spec Kit (inferred)
These agents are listed in specify init --ai <name> but haven't
been hands-on tested with speckit-security. They should work
because Spec Kit handles the translation layer and the extension
provides the same agent-neutral Markdown that it translates for the
verified agents above.
auggiebobcodebuddycodexforgeiflowjuniekilocodekimikiro-clipiqodercliqwenrooshaitabninetraevibe(Mistral Vibe)windsurfgeneric(fallback for unsupported agents)
If you try one of these and it works, please open a PR upstream to
move it into the "Verified" table. If it doesn't work, please open
an issue with the exact specify init and specify extension add
output.
Requirements regardless of agent
- Spec Kit
>= 0.1.0installed viauv tool install specify-cli - Bash (macOS + Linux supported; Windows requires WSL or Git Bash)
- Python 3, used by helper scripts for JSON parsing and template substitution
Testing on your agent of choice
# 1. Init a project with your target agent
specify init test-app --ai <agent-name> --no-git
# 2. Install the extension
cd test-app
specify extension add --dev /path/to/speckit-security
# 3. Verify registration
specify extension list
# Should show: Commands: 8 | Hooks: 5 | Status: Enabled
# 4. Verify the commands were translated into the agent's format
specify extension info tekimax-security
# 5. Run the gate-check script directly against a sample spec
# (it requires a real spec path, so create a stub first)
mkdir -p .specify/specs
cat > .specify/specs/F-000-smoke.md <<'EOF'
## Data Contract
## Security / Threat Model
EOF
bash .specify/extensions/tekimax-security/scripts/bash/gate-check.sh \
.specify/specs/F-000-smoke.md 2>&1 || trueIf step 3 shows Commands: 8 | Hooks: 5 | Status: Enabled and
step 5 runs the gate script (even if it reports gate failures on
the stub spec), the agent is functionally supported.
Reporting incompatibilities
- Open an issue at https://github.com/TEKIMAX/speckit-security/issues
- Include the agent name, spec-kit version, OS, and the exact
output of
specify extension add --devandspecify extension list - Do not include prompts or conversation snippets, just the commands you ran and their output