codex-sdk — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited codex-sdk (Agent Skill) and scored it 91/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 1 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 1 flagged
The text {match} is the classic direct prompt-injection phrasing. Placed in a skill body that the agent reads as trusted instructions, it tries to make the agent abandon its prior rules and follow whatever comes next — a full system-prompt override.
ignore/disregard/forget … previous instructions sentence.Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Build reliable, auditable, multi-step coding workflows that scale from a single run to multi-agent orchestration.
For multi-hour work, keep intent durable across compaction/restarts using an ExecPlan:
references/execplans.md.agent/PLANS.md (use scripts/init_agent_workspace.py to bootstrap a repo)assets/templates/execplan.md (or generate a file with scripts/new_execplan.py)codex exec (references/codex-cli-exec.md)@openai/codex-sdk (references/codex-sdk-typescript.md)codex mcp-server and orchestrate via OpenAI Agents SDK (references/mcp-and-agents-sdk.md)references/state-memory-sqlite.md)SQLite is the simplest reliable substrate for:
codex exec --json event as an immutable audit logthreadId so runs can resume deterministicallypython3 scripts/codex_jsonl_to_sqlite.py --db codex-runs.sqlite --initcodex exec --json "<prompt>" | python3 scripts/codex_jsonl_to_sqlite.py --db codex-runs.sqlite --run-label "ci-autofix"python3 scripts/codex_sqlite_report.py --db codex-runs.sqlite --latestUse:
Details: references/mcp-and-agents-sdk.md
Prefer:
sandbox: read-only, approval-policy: neversandbox: workspace-write, approval-policy: on-request / on-failureexecpolicy rules (references/safety-and-execpolicy.md)references/codex-sdk-typescript.md – SDK patterns (threads, streaming, schemas)references/codex-cli-exec.md – CLI patterns (JSONL, schema files, resume)references/mcp-and-agents-sdk.md – Codex as MCP server + multi-agent orchestrationreferences/agents-sdk-consistent-workflows.md – gated handoffs + traces with Codex MCP + Agents SDKreferences/execplans.md – ExecPlans for long-running work across compactionreferences/state-memory-sqlite.md – SQLite schema + memory/caching patternsreferences/safety-and-execpolicy.md – sandboxing, approvals, prompt-injection defensesreferences/codex-config-knobs.md – config keys and feature flags that matterreferences/orchestration-patterns.md – planner/executor/verifier and orchestrator/worker patternsreferences/rag-and-memory.md – SQLite-first shared memory and RAG guidancereferences/context-personalization.md – state + memory notes personalization patterns (Agents SDK)scripts/codex_jsonl_to_sqlite.py – ingest Codex JSONL into SQLitescripts/codex_sqlite_report.py – summarize runs from SQLitescripts/init_agent_workspace.py – create .agent/AGENTS.md + .agent/PLANS.md from templatesscripts/new_execplan.py – generate execplans/execplan-*.md from the ExecPlan templateassets/templates/ – copy/paste templates (ExecPlan, prompts, schemas)assets/templates/agents-sdk/ – Agents SDK starter snippets (MCP stdio, sessions, personalization)~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.