# Codex residual mixed-runtime marker allowlist
#
# Scope: skills-codex/**/SKILL.md
#
# Format:
# - One POSIX ERE marker pattern per line.
# - Matching is case-sensitive.
# - Lines beginning with "#" are comments.
#
# Policy:
# - Keep entries exact and minimal.
# - Only allow markers required for intentional mixed-runtime behavior.
# - Do not add broad catch-alls (for example: "Claude", "claude", ".*", or ".*claude.*").
# - Prefer stable contract identifiers (filenames, backend IDs, primitive names).

# Cross-vendor mode label (explicit mixed-vendor behavior)
\bClaude \+ Codex\b

# Runtime backend naming (Claude fallback contract)
\bClaude Native Teams\b
\bclaude-native-teams\b
\bclaude_teams\b

# Shared backend contract references
\bbackend-claude-teams\.md\b
\bclaude-cli-verified-commands\.md\b
\bclaude-code-latest-features\.md\b

# Claude-native team primitives referenced by runtime-neutral orchestration docs
\bteam-create\b
\bsend-message\b

# Intentional cross-runtime comparisons in codex skills
[Cc]laude[^[:cntrl:]]*[Cc]odex|[Cc]odex[^[:cntrl:]]*[Cc]laude
Codex CLI processes as background shell commands
ao overnight setup --apply --runner codex --runner claude --at 01:30
ao overnight start --goal "stabilize release follow-ups" --runner codex --runner claude --creative-lane

# Explicit alternate-runtime backend references
\bClaude teams\b
\bClaude runtime\b
\bclaude agents\b
\bCOUNCIL_CLAUDE_MODEL\b
council/claude-\*
\.claude-plugin/plugin\.json

# Agent count descriptions in council/shared flags tables
\bClaude agents instead of\b
\b[0-9]+ Claude\b

# Cross-runtime tracking comparisons (beads vs TaskList)
\bClaude-native\b
\bTaskList \(Claude-native\)
\bClaude uses [0-9]+ tools\b

# Runtime fallback command references
\bcommand -v claude\b
\bruntime_command: claude\b
\bClaude custom agents\b

# Domain checklist trigger patterns containing glob syntax (not skill invocations)
\*llm\*
\*prompt\*
\*completion\*
Manifest versions sync
anthropic`, `openai` imports
`anthropic`, `openai`, `google\.generativeai`
