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 skills add AAWWCC/ape --skill rungit clone --depth 1 https://github.com/AAWWCC/apeWrote 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/aawwcc/ape/run)<a href="https://agentmods.dev/skills/aawwcc/ape/run"><img src="https://agentmods.dev/badge/skills/aawwcc/ape/run/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/aawwcc/ape/run"><img src="https://agentmods.dev/badge/skills/aawwcc/ape/run.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00013 | $0.00834 |
| Opus 5.5 | $0.00005 | $0.00334 |
| Sonnet 5.5 | $0.00003 | $0.00167 |
| Haiku 4.5 | $0.00001 | $0.00083 |
Grade A, and why
run 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 1mo 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.
What it actually says
APE run
Use only when the user explicitly asks to run APE; never infer consent.
Use applicable repository instructions already supplied by the host. If further discovery is
needed, include AGENTS.override.md as well as AGENTS.md, including ignored local files.
These files are optional: no match is valid and does not block intake.
objective: outcome and acceptance. Omit execution budgets and dispatch limits; new workers have no duration limit, and preview reportsdeadline_ms: null.host(codexorclaude) in preview/start.- Include confirmed
hooks_trusted: true,subagents_available: true, andexplicit_invocation: truein preview and unchanged start. Never invent trust or availability. Host invocation policy is the human-intent boundary;explicit_invocation: trueis caller-attested defense-in-depth, not proof of human intent. mode:phase,debug,spike, orland;lane:autounless explicitly selected.claimed_paths: production paths only. Include generated artifacts or documentation only when the objective may require them.test_paths: independently authored test paths; never put them inclaimed_paths.test_intent:red-first, or explicitgreen-maintenancefor green-on-arrival coverage/deflaking. Nonbehavioral data/baseline work usesbehavioral: false. Keep preview/start identical.behavioral,requirements,completes, andrisk_triggers.required_capabilities: exact extra command/verification IDs. External tools remain host-owned.run_command_profiles:debug/spikeonly; declare command, read-only role,effect: execute, and reason. Setoperator_authorized: trueonly after explicit approval.plan_contract_version: 2for every newly started behavioral fast/fullphase; omit it for mechanical work, non-phase modes, and every resume. Version 1 is legacy-only.
Call ape_config doctor/get, then ape_run preview. Follow the shared protocol's start-readiness
steps for missing configuration, complete inline or paged admission review, and the binding probe.
Start only after reviewing the complete ready manifest, with unchanged prospective inputs and
expected_admission_digest copied from the reviewed preview or paged delivery. Changed inputs
require a fresh preview. Report deterministic
dispatch bounds and complete gate-command and visual-evidence readiness checks.
One explicit APE invocation authorizes the run. When shipping.auto_merge is true, runtime
freezes shipping authority; omit legacy auto_merge_authorized. Drive every scheduler-owned
transition, wait, and configured shipping action to a terminal result without asking for continue.
Shipping requires the explicitly configured project target and passes that frozen
target to every mutation; the canonical APE checkout separately retains its public-repository guard.
Ask only unresolved outcome-changing decisions; existing explicit approval remains valid within scope. Offer a roadmap for cross-run work; registration requires explicit approval. Decompose independent high-risk subsystems per the protocol.
Follow references/run-resume-protocol.md for every action.
During the run, the parent owns control calls and stage workers edit production/tests;
the runtime owns sequencing, retries, remediation, gates, and shipping.
When preflight returns input_required, collect complete exact answers for every question id. Call
ape_run answer-preflight with exact run/hash, bounded audit reason, and only additive
claimed_paths, test_paths, and canonical risk_triggers; never subtract/reinterpret scope.
What ships with it
1 file 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.
- 1mo ago Changed · -6 lines b48ca53f67c4
- 1mo ago Changed c35c609a87aa
- 2mo ago Changed · -3 lines bf5de003cf0b
- 2mo ago Changed · +1 lines f4307c653a9e
- 2mo ago First seen · 68 lines · 13 tokens per session scan A 598ce057396e
run is a skill published in the GitHub repository AAWWCC/ape (0 stars, last pushed 24d ago), licensed MIT. It adds 13 tokens to every session and 834 once invoked, about $0.0001 per session on Opus 5.5. 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.
run compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| runAAWWCC/ape | Skill | 0 | 24d ago | MIT | A | 13 | Claude Code |
| test-writingandreaswasita/copilot-agents-dojo | Skill | 55 | 2mo ago | MIT | A | 11 | Claude Code, Codex |
| tddSetraTheXX/Codex-Engineering-Workflow-Pack | Skill | 1 | 1mo ago | MIT | A | 43 | Claude Code, Codex |
| django-tddhashgraph-online/awesome-codex-plugins | Skill | 1.6k | today | Apache-2.0 | A | 32 | Claude Code, Codex |
| laravel-tddhashgraph-online/awesome-codex-plugins | Skill | 1.6k | today | Apache-2.0 | A | 27 | Claude Code, Codex |
| project-executionathola/claude-night-market | Skill | 350 | yesterday | MIT | A | 31 | Claude Code |
| iterative-developmentalinaqi/maggy | Skill | 711 | 1mo ago | MIT | A | 24 | Claude Code |