oracle-codex
npx skills add https://github.com/paulrberg/dot-agents --skill oracle-codex
Agent 安装分布
Skill 文档
Codex Oracle
Use OpenAI Codex CLI as a read-only oracle â planning, review, and analysis only. Codex provides its perspective; you synthesize and present results to the user.
Sandbox is always read-only. Codex must never implement changes.
Arguments
Parse $ARGUMENTS for:
- query â the main question or task (everything not a flag). Required â if empty, tell the user to provide a query and stop.
--reasoning <level>â override reasoning effort (low,medium,high,xhigh). Optional; default is auto-selected based on complexity.
Prerequisites
Run the check script before any Codex invocation:
scripts/check-codex.sh
If it exits non-zero, display the error and stop. Use the wrapper for all codex exec calls:
scripts/run-codex-exec.sh
Configuration
| Setting | Default | Override |
|---|---|---|
| Model | gpt-5.3-codex |
Allowlist only (see references/codex-flags.md) |
| Reasoning | Auto | --reasoning <level> or user prose |
| Sandbox | read-only |
Not overridable |
Reasoning Effort
| Complexity | Effort | Timeout | Criteria |
|---|---|---|---|
| Simple | low |
300000ms | <3 files, quick question |
| Moderate | medium |
300000ms | 3â10 files, focused analysis |
| Complex | high |
600000ms | Multi-module, architectural thinking |
| Maximum | xhigh |
600000ms | Full codebase, critical decisions |
For xhigh tasks that may exceed 10 minutes, use run_in_background: true on the Bash tool and set CODEX_OUTPUT so you can read the output later.
See references/codex-flags.md for full flag documentation.
Workflow
1. Parse and Validate
- Parse
$ARGUMENTSfor query and--reasoning - Run
scripts/check-codex.shâ abort on failure - Assess complexity to select reasoning effort (unless overridden)
2. Construct Prompt
Build a focused prompt from the user’s query and any relevant context (diffs, file contents, prior conversation). Keep it direct â state what you want Codex to analyze and what kind of output you need. Do not implement; request analysis and recommendations only.
3. Execute
Invoke via the wrapper with HEREDOC. Set the Bash tool timeout per the reasoning effort table above.
EFFORT="<effort>" \
CODEX_OUTPUT="/tmp/codex-${RANDOM}${RANDOM}.txt" \
scripts/run-codex-exec.sh <<'EOF'
[constructed prompt]
EOF
For xhigh, consider run_in_background: true on the Bash tool call, then read CODEX_OUTPUT when done.
4. Present Results
Read the output file and present with attribution:
## Codex Analysis
[Codex output â summarize if >200 lines]
---
Model: gpt-5.3-codex | Reasoning: [effort level]
Synthesize key insights and actionable items for the user.