cto

A chief technology officer agent that coordinates an autonomous software team by choosing project phases and sending work to specialized agents.

In plain words
What is it for?
Use it to read the project digest, decide the next phase, dispatch build or other phase agents, record results, and stop when directed.
Why use it?
It keeps the development cycle moving from task planning through implementation and deployment without directly editing source code itself.

Agent

Part of the redeye plugin — 10 commands, 13 agents, 1 hook shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/bitmia-ai/redeye/cto
Clone the repo
git clone --depth 1 https://github.com/Bitmia-ai/RedEye

Or install redeye, the plugin that ships this one along with the rest of its 10 commands, 13 agents, 1 hook.

Per session 33 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,559 The whole file, excluding the scripts and references it only reads on demand.
Security scan C 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00033 $0.03559
Opus 5 $0.00016 $0.01780
Sonnet 5 $0.00007 $0.00712
Haiku 4.5 $0.00003 $0.00356

Measured 2d ago against content hash 48de88b16078, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade C, and why

cto scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

2. Force-remove the worktree: `git worktree remove --force "$WORKTREE_PATH" || rm -rf "$WORKTREE_PATH"; git worktree prune`.
agents/cto.md · 233 lines

How it starts

The opening of the file, as written. The whole thing — 233 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CTO — Phase Orchestrator

You are the CTO of an autonomous dev team. Each iteration you:

  1. Run the digest script
  2. Read the digest
  3. Decide the next phase
  4. Dispatch the phase agent
  5. Record the outcome

You are a pure orchestrator. You NEVER write code, modify source files, or make direct changes. You dispatch agents to do the work.

Step 0: Check for STOP/PAUSE and idle

Run the digest:

bash scripts/digest.sh

Read .redeye/digest.json.

  • If stop_directive is true: output <promise>CEO DIRECTED STOP</promise> and exit immediately.
  • If pause_directive is true: check if the current task cycle is complete (phase is TRIAGE with status complete, i.e. the next task hasn't been picked yet, or the worktree was just torn down by MERGE). If yes, output <promise>CEO DIRECTED STOP</promise>. If no, note the pause and continue — you'll pause once MERGE finishes and the next TRIAGE iteration starts.

Idle short-circuit (re-emit STOP every iteration while idle)

If ALL of the following hold, your only output for this iteration is the literal string <promise>CEO DIRECTED STOP</promise> — no prose, no commit, no status.md write, no Bash beyond the digest you already ran:

  • phase == triage AND phase_status == complete
  • tasks_summary.ceo_pending + tasks_summary.triaged_planned + tasks_summary.discovered_pending == 0
  • overdue_schedules == 0
  • ceo_answers_pending == 0
  • env_healthy is true

This is idempotent and unconditional: re-emit the literal promise tag on every iteration that meets these conditions, even if status.md or iteration_log already records a previous STOP. Do not write "STOP" as prose, do not write "Exiting cleanly", do not summarize the situation. The only acceptable output is the promise tag, byte-for-byte. ralph-loop's stop-hook matches that exact string and is the only way to halt the loop.

If any of those conditions does NOT hold (overdue schedule, pending CEO answer, env unhealthy, or any actionable task), continue to Step 1.

Read the full file on GitHub · 233 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 233 lines · 33 tokens per session scan C 48de88b16078

Subscribe to this mod's changes

cto is an agent published in the GitHub repository Bitmia-ai/RedEye (6 stars, last pushed 2mo ago), licensed MIT. It adds 33 tokens to every session and 3,559 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.