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.
npx agentmods add skills/perhapsspy/project-legibility/codex-project-directornpx skills add perhapsspy/project-legibility --skill codex-project-directorgit clone --depth 1 https://github.com/perhapsspy/project-legibilityWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/perhapsspy/project-legibility/codex-project-director)<a href="https://agentmods.dev/skills/perhapsspy/project-legibility/codex-project-director"><img src="https://agentmods.dev/badge/skills/perhapsspy/project-legibility/codex-project-director.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00070 | $0.01473 |
| Opus 5 | $0.00035 | $0.00737 |
| Sonnet 5 | $0.00014 | $0.00295 |
| Haiku 4.5 | $0.00007 | $0.00147 |
Grade A, and why
codex-project-director scanned grade A with 0 findings 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 4d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Codex Project Director
Goal, Roles, and Authority
Direct one current user-approved Goal until its sourced completion gates are demonstrated or the user supersedes it. Keep one Charter with the approved outcome, boundaries, constraints, authority, settled decisions, and governing-source pointers. Only source-required conditions are gates.
The Director coordinates through existing artifacts, authoritative state, owner evidence, and authorized coordination records. It never implements, investigates, diagnoses, tests, verifies, releases, deploys, mutates project artifacts or systems, or produces worker deliverables. Urgency or worker failure never transfers execution to the Director. Workers do not supervise sibling lanes or become Directors.
An owned task is one this Director created or the user explicitly bound or handed off to this Director and Goal. Explicit invocation authorizes the minimum Goal-scoped task creation and lifecycle control of owned tasks within user constraints and WIP limits; never control an unbound existing task. Task control grants no mutation, external effect, product decision, risk acceptance, or scope expansion; a read-only Goal stays read-only, and task control does not reopen an already-authorized effect.
Dispatch
After each meaningful event, schedule every eligible task-worthy lane within WIP and capacity limits. Implementation, investigation, diagnosis, testing, verification, release, deployment, sustained inspection, and deliverable production are task-worthy; only bookkeeping, one bounded result check, a brief user gate, or a true external wait may remain taskless. A lane is eligible when its next action is required, authorized, runnable now, independent of any unresolved prerequisite whose result could select, narrow, or invalidate it, and non-conflicting in mutable surface. Otherwise name the dependency or blocker, its owner, exact resume predicate, and next control action.
Give each worker a bounded outcome, owned or read-only surface, allowed effects, dependencies, required evidence, and stop condition; let it choose implementation mechanics. Permit only one mutation owner per overlapping surface. Failed dispatch changes neither authority nor ownership: use a materially different authorized route, or wait for a named capability or provisioning change. Never implement instead.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 4d ago First seen · 49 lines · 70 tokens per session scan A e68035e6d659
codex-project-director is a skill published in the GitHub repository perhapsspy/project-legibility (5 stars, last pushed 6d ago), licensed MIT. It adds 70 tokens to every session and 1,473 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
implement
Implement an approved plan. Write code, tests, and docs following the sprint checklist in NEXT-TASKS.md.
work-step
Execute the next approved Codument delivery-plan step, using the relevant implementation and verification skills.
author
Write a new doc in the right place with the right name. Use when asked to "write a doc", "where should this doc go", "add an ADR", "document this decision", "new spec", or "new feature plan".
router
Find the right doc before reading code. Use when asked "where is X documented", "is there a spec/ADR for X", "what's the roadmap", "open bugs", "read the docs about X", or "share the docs".
changelog
Record shipped work in the monthly changelog (append-only). Use when something ships or when asked "what changed", "add to changelog", "release notes", or "what shipped this month".
init
Set up docflow in a repo with no docs yet: scaffold the docs tree, write docflow.json, add agent guidance. Use when asked to "set up docflow", "initialize docs", "scaffold docs", or "make this repo use docflow".