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 agents/thuandiep1707/codex-harness-javascript/subagentsgit clone --depth 1 https://github.com/thuandiep1707/codex-harness-javascriptWhat 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 | $0.00000 | $0.01179 |
| Opus 5 | $0.00000 | $0.00589 |
| Sonnet 5 | $0.00000 | $0.00236 |
| Haiku 4.5 | $0.00000 | $0.00118 |
Grade A, and why
subagents 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 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.
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 — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Agent runtime map
Public workflow entry points
| Workflow | Execution intent | Result |
|---|---|---|
$frontend-delivery |
deliver |
Continue end-to-end until acceptance/pause/blocker |
$frontend-planning |
plan-only |
Create/reconcile Jira work graph then stop |
Lifecycle planning/resume/pause is separate from execution intent. Planning mode does not imply plan-only.
Runtime roles
| Role | Codex agent | Responsibility |
|---|---|---|
| Primary controller | main chat | Resolve workflow/lifecycle; own native child-agent lifecycle, Jira connector transport, and runtime cleanup supervision; execute Orchestrator controller actions |
| Brain | brain |
Analyze relevant .docs, detect implementation environment, record baseline, targeted revalidation, final acceptance |
| Orchestrator | orchestrator |
Jira/workflow decisions, capability routing, specialist dispatch decisions, resume, pause/handoff decisions, reconciliation |
| Design | design |
Use connected design provider and return design evidence |
| Test plan | test-plan |
Produce a bounded risk-based test-plan result |
| Coding | coding |
Implement one bounded Coding Subtask using routed internal capabilities |
| Testing | testing |
Implement/run one bounded Testing Subtask using routed internal capabilities |
The Primary Controller is intentionally thin but is the only runtime transport owner. Brain does not create tasks. Orchestrator does not perform specialist implementation, invoke native child-agent lifecycle APIs, or call Jira directly. Chat history is never required to resume valid Jira work.
Controller action loop
Orchestrator makes the decision; Primary Controller performs the transport:
Orchestrator
→ controller-action (jira-call | dispatch-specialist)
→ Primary Controller executes exactly
→ confirmed controller-action-result
→ same Orchestrator child
→ next action or terminal result
Keep one Orchestrator child alive across an active workflow instead of respawning it after every specialist result. Specialists are short-lived sibling children spawned and closed by the Primary Controller.
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.
- 2d ago First seen · 123 lines · 0 tokens per session scan A 34de05bd4a5e
subagents is an agent published in the GitHub repository thuandiep1707/codex-harness-javascript (7 stars, last pushed 4d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 1,179 tokens. 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 agents, from other repositories
independent-reviewer
Fresh-context read-only independent reviewer for SPEC-based acceptance evidence and integration recommendation.
task-specialist
Bounded implementer for one ready task with an exclusive write set and approved context.
discovery-interviewer
Read-only interviewer for missing, stale, or mature-project context before implementation planning.
learning-assessor
Read-only project-learning assessor used only after explicit learning activation and observation consent.
ap-implementer
L3 executor - G4 IMPLEMENT. Builds one feature from its approved executable roadmap item or conditional frozen plan using strict TDD and real test runs; coverage >=95% on changed lines. Reports PLAN-CONFLICT rather than improvising.
ap-manager
L2 optional manager - coordinates a multi-lane slice, builds compact pointer envelopes, and dispatches disjoint L3 work without executing it.