| Version | Supported |
|---|---|
| 1.0.x | ✅ |
ultrapilot is a skill that runs locally in your coding agent. It is not a service, server, or hosted product.
- All state is stored in
~/.ultrapilot/goals.db(a SQLite file on your local machine) - All scripts (
scripts/ultrapilot_goals.py,scripts/ultrapilot_run.py) run as subprocesses of your agent - No network calls are made by ultrapilot itself
- The skill reads files in your project, the goal DB, and the agent's environment
ultrapilot's threat model is:
-
Prompt injection via the goal text — mitigated by the
<objective>wrapper incommands/goals.mdand theultrapilot_goals.py suggestprompt renderer. The goal text is treated as untrusted task context, not as instructions. -
State injection via DB tampering — mitigated by the source-attribution field (
agentcolumn) and the fact that ultrapilot's own scripts are the only writers. A malicious actor with file-system access can tamper with the DB; that is the same threat as any local file. -
Runaway resource usage — mitigated by:
- The 500-continuation runaway guard (
ULTRAPILOT_MAX_CONTINUATIONS) - The token budget soft bound (
--tokens N) - The phase-state machine that prevents the agent from looping outside the intended flow
- The 500-continuation runaway guard (
-
Cross-session leakage — mitigated by multi-source session ID resolution. Goals do not leak across separate sessions.
-
Goal-text prompt-injection hijacking the agent — the agent is told explicitly to treat
<objective>content as untrusted and to ignore instructions that contradict higher-priority context.
If you find a security issue in ultrapilot:
- Email: create a GitHub issue marked
securityor email the maintainers. - Do not post the vulnerability publicly until it has been triaged.
- We aim to respond within 72 hours.
- Bugs in the agent's behavior (use bug report template)
- Documentation issues
- Feature requests
- Performance issues
We follow responsible disclosure:
- Reporter notifies us privately.
- We confirm and triage within 72 hours.
- We develop a fix and a CVE if appropriate.
- We coordinate disclosure timing with the reporter.
- We publish a security advisory after the fix is released.
- We will not silently fix vulnerabilities.
- We will not blame reporters for honest mistakes.
- We will not pursue legal action against good-faith security research.