spar-cli 0.1.10

Two AI coding agents alternate implementing and reviewing GitHub issues until a PR converges.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34

# OpenAI Codex. Ships inside the ChatGPT desktop app on macOS, or standalone.
#
# Note: `codex exec review` is deliberately not used. It refuses a custom prompt
# alongside --base and returns prose regardless of --output-schema, so it cannot
# produce a machine-checkable verdict.
command = [
  "codex", "exec", "--json", "--skip-git-repo-check",
  ["-s", "workspace-write"],
  ["-m", "{model}"],
  ["-c", "model_reasoning_effort={effort}"],
  ["-C", "{cwd}"],
  ["--output-schema", "{schema_file}"],
  "--", "{prompt}",
]
output = "jsonl"
message_path = "item.text"
search_paths = [
  "/Applications/ChatGPT.app/Contents/Resources",
  "~/.codex/bin",
  "/opt/homebrew/bin",
  "/usr/local/bin",
  "~/.local/bin",
]

# Hints for `spar init`. Never validated against, only written as comments. The
# Codex CLI does not enumerate these in its help, so this lists what is known to
# work rather than guessing at a full set.
models = ["gpt-5.6-sol"]
efforts = ["ultra"]
options_note = "Codex does not list these in its help. Your ~/.codex/config.toml is the source of truth, and effort is passed through as -c model_reasoning_effort=..."

[message_match]
type = "item.completed"
"item.type" = "agent_message"