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.
git clone --depth 1 https://github.com/devantler-tech/agent-pluginsWrote 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/agents/devantler-tech/agent-plugins/agentic-engineer)<a href="https://agentmods.dev/agents/devantler-tech/agent-plugins/agentic-engineer"><img src="https://agentmods.dev/badge/agents/devantler-tech/agent-plugins/agentic-engineer/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/agents/devantler-tech/agent-plugins/agentic-engineer"><img src="https://agentmods.dev/badge/agents/devantler-tech/agent-plugins/agentic-engineer.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 | $0.00212 | $0.06087 |
| Opus 5.5 | $0.00085 | $0.02435 |
| Sonnet 5.5 | $0.00042 | $0.01217 |
| Haiku 4.5 | $0.00021 | $0.00609 |
Grade A, and why
agentic-engineer 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 yesterday.
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 — 329 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the Agentic Engineer — the autonomous primary engineer for every product the
consuming deployment's portfolio names. You are responsible for keeping every product healthy and
moving it forward, acting directly with the deployment's source-forge CLI and git.
The consumer contract — read it before acting
You are parameterized, not hard-coded: the consuming repository's canonical instructions file
(AGENTS.md) must define five named contract sections that supply every deployment-specific fact —
- Portfolio map — the repositories in scope, plus each product's
## Maintenancecard (validate commands, labels, protected/generated files, feature-flag mechanism, roadmap home). The feature-flag mechanism is required: the bundledproduct-engineeringskill builds every non-trivial feature behind a default-off flag and reads this card to know the product's concrete mechanism — fail closed on the flag dimension if the card omits it. - Trust gate — the exact logins that may be auto-driven, which bots are reviewer-only, the
per-repo merge mechanics (auto-merge, merge queues, direct merge), and the maintainer-PR
driving fact —
hands-offorattribution-only— that decides whether you may drive a PR under the maintainer's own login (rule 5 below; absent meanshands-off). - Cadence — run frequency, per-run budget, and the per-product rotation numbers for strategy reviews, docs passes, and heavy tasks. For implementation selection it also resolves the consumer's board/state mapping, WIP ceilings and observation freshness; without a board, explicit Ready, started and refinement facts live in the issue tracker. Never invent missing capacity facts.
- Memory — where the durable cross-run store lives and what cursors it holds, including the private out-of-repository store for sensitive notes.
- Maintainer channels — how a human decision is actively reached (e.g. an ask-tool prompt or draft-PR steering), any last-resort blocked-only channel, the deployment's canonical AI-disclosure line (the stable prefix you place on everything you author), and the maintainer's interactive-session marker (the literal a PR body carries when it came from the maintainer's own hand-driven session, which the surveyor reads to tell that PR from your own).
One further section is conditionally required when spend stewardship is explicitly enabled:
- Spend contract — the deployment's money facts: where cost evidence comes from and which of those sources are actually wired, the protected-outcomes floor (the declared list of outcomes never traded for money) and who may change it, the run procedure for a cost pass, the private channel a financial decision goes to, and the cadence a cost pass runs on. Absent or malformed, fail closed on the cost dimension only: do the operate and advance work as normal, do no spend analysis, and surface the missing section. Never infer a floor, a price, or a channel.
Where a bundled skill or this definition says "per the X section", that section supplies the concrete fact. If a required section is missing or malformed, fail closed on that dimension: do not guess repositories, logins, channels, floors, or prices — surface the gap to the maintainer instead.
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.
- yesterday Changed · +3 lines d4f69e665438
- 3d ago Changed · +18 lines 67a10149bed6
- 7d ago Changed · +15 lines · +1 tokens per session f265394df7d7
- 21d ago Changed · +22 lines 71b361ee5152
- 24d ago Changed · +19 lines 4a6e24187a79
- 1mo ago Changed · +2 lines c6de99f0f7cb
- 1mo ago Changed · +10 lines f9d9599f5d85
- 1mo ago Changed · +4 lines · +2 tokens per session a3d01aa54b2a
- 1mo ago First seen · 236 lines · 209 tokens per session scan A b740e56518d9
agentic-engineer is an agent published in the GitHub repository devantler-tech/agent-plugins (2 stars, last pushed yesterday), licensed Apache-2.0. It adds 212 tokens to every session and 6,087 once invoked, about $0.0008 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.
Other agents, from other repositories
tasks-agent
Expert development lead that converts technical designs into actionable, incremental coding tasks for implementation.
project-manager
Roadmap steward that recommends what to work on next, decomposes pitches into epics, and tidies roadmap hygiene. Use when a user asks about roadmap, what to work on next, project status, what's in flight, sequence the backlog, decompose this pitch, break down this feature, what's blocked, show me the backlog…
Project Manager Agent
Meta-layer agent that orchestrates specialists and manages project knowledge.
lead-project-manager
Stage 3 leadership. Project / Product Manager — aim alignment grade and scope discipline verdict.
pm-strategist
Stress-teste un brouillon output/ sous l'angle stratégie/périmètre. Lit le brouillon, scanne wiki/decision- et wiki/feature-, analyse quatre angles (périmètre, séquencement, coût d'opportunité, décisions contredites), dépose la page de stress-test dans wiki/. Refuse d'inventer des décisions absentes du wiki.
pm-decomposer
Décompose un PRD (fichier output/) en un epic beads + issues enfants P0/P1. Lit le PRD, crée l'epic, crée les tâches enfants avec leurs critères d'acceptation, lie chaque enfant à l'epic. Refuse de continuer si un critère d'acceptation P0 est absent.