Security
Destructive operations, authentication, secret handling, and prompt injection.
Failure modes: 5 active | 🔴 5 critical
| File | Severity | Summary |
|---|---|---|
| 23-critical-destructive-ops.md | 🔴 Critical | Agents may execute destructive commands without fully understanding consequences, especially when operating autonomou... |
| 24-critical-auth-secrets.md | 🔴 Critical | CLI tools often need credentials |
| 25-critical-prompt-injection.md | 🔴 Critical | CLI tool output is fed directly into the agent's context |
| 74-critical-credential-scope-declaration.md | 🔴 Critical | CLI tools accept any credential without declaring what permissions each command requires, giving agents full account blast radius |
| 75-critical-safe-default-execution.md | 🔴 Critical | High-stakes commands execute immediately with real side effects; agents have no way to preview impact without explicitly remembering --dry-run at every callsite |
Detailed Metrics
| Challenge | Severity | Frequency | Detectability | Token Spend | Time | Context |
|---|---|---|---|---|---|---|
| §23 | 🔴 Critical | Common | Medium | Medium | High | Medium |
| §24 | 🔴 Critical | Common | Hard | Medium | Medium | Low |
| §25 | 🔴 Critical | Situational | Hard | High | High | High |
| §74 | 🔴 Critical | Common | Hard | Low | Medium | Low |
| §75 | 🔴 Critical | Situational | Hard | Low | Critical | Low |