Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/AndreVianna/aid-methodologynpx agentmods add skills/andrevianna/aid-methodology/aid-executeWrote 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/andrevianna/aid-methodology/aid-execute)<a href="https://agentmods.dev/skills/andrevianna/aid-methodology/aid-execute"><img src="https://agentmods.dev/badge/skills/andrevianna/aid-methodology/aid-execute/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/andrevianna/aid-methodology/aid-execute"><img src="https://agentmods.dev/badge/skills/andrevianna/aid-methodology/aid-execute.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00095 | $0.05690 |
| Opus 5 | $0.00048 | $0.02845 |
| Sonnet 5 | $0.00019 | $0.01138 |
| Haiku 4.5 | $0.00010 | $0.00569 |
Grade A, and why
aid-execute 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 8d 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 — 440 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Execute Task
Read the type. Do the work. Review it. Fix it. Ship it.
⚠️⚠️ MANDATORY: write task State AS IT CHANGES -- read this first
Whoever executes this task -- the main/orchestrator agent running it DIRECTLY, or a dispatched sub-agent -- MUST write the task's
Statefield the instant it changes:In Progressbefore starting work,In Reviewbefore dispatching the reviewer, and a terminal value (Done/Failed) when finished. No agent may skip, batch, or defer these writes, on the flat OR full layout, whether this task is run alone or as part of a pool dispatch. The full mandate, the exact command for each transition, and the why lives inreferences/state-execute.md § MANDATORY: State-Write Protocol-- read it before Step 1 below. Skipping these writes is the single most common root cause of a task that showsPendingin the dashboard for its entire execution and then jumps straight toDone.
⚠️ Pre-flight Checks
Check 1: Locate Work and Task
-
Read first arg: if it starts with
work-→ use that work directory; if it starts withtask-→ treat as shorthand (single-work auto-select below) -
If work arg not provided (or shorthand): enumerate works cross-worktree: run
bash .claude/aid/scripts/works/enumerate-works.sh(main tree + every git worktree; never the local.aid/works/glob, which is empty onmaster), taking each record's field-1work_id— single record → auto-select; multiple records → list them, ask user to choose; zero records on any worktree → STOP. "No works found. Run/aid-describefirst." -
Read second arg (or first arg when shorthand): the
task-NNNidentifier; also resolvedelivery-NNNfrom the task identifier or from the Source fieldLocate + enter the work's worktree. Now that the work id is resolved (steps 1-3 above) and before step 4 below reads anything under
.aid/works/{work}/…, follow.claude/aid/templates/downstream-worktree-entry.mdto normalize the work id to its barework-NNNbranch name,locatethe worktree (which always exits 0 and returns<path>\t<status>), and enter the returned path. Keep the defensive empty-path/non-zero backstop that stops rather than operate blindly — it should not fire against the real helper. Never create a new worktree — creation belongs to the work-starting skills only. -
Detect the layout (per-work, DECLARED first; additive — does not change the full-path resolution below). Read
pipeline.pathfrom the work-rootSTATE.yml:lite→ flat,full→ nested. Infer from file presence ONLY when no value is declared, which means an un-migrated work:- Full path (nested):
.aid/works/{work}/deliveries/exists → tasks live atdeliveries/delivery-NNN/tasks/task-NNN/DETAIL.md. - Flat path (feature-001, single-delivery):
.aid/works/{work}/tasks/task-NNN/DETAIL.mdexists directly under the work root AND nodeliveries/wrapper under it → the delivery is always the synthesizeddelivery-001(no per-taskSTATE.yml; task mutable cells live in the work-rootSTATE.yml'stasks_lifecyclemapping instead — see## Workspacebelow).
- Full path (nested):
-
Find the task definition at the layout-appropriate path from step 4.
-
Task not found → STOP. List available tasks.
What ships with it
9 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.
- references/reviewer-brief.md 7.6 KB
- references/reviewer-guide.md 3.4 KB
- references/state-delivery-gate.md 22 KB
- references/state-execute-drilldown.md 7.7 KB
- references/state-execute.md 42 KB
- references/state-fix.md 9.3 KB
- references/state-re-run.md 1.1 KB
- references/state-review.md 9.2 KB
- references/task-type-rules.md 3.9 KB
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.
- 8d ago First seen · 440 lines · 95 tokens per session scan A 62ee774793e6
aid-execute is a skill published in the GitHub repository AndreVianna/aid-methodology (5 stars, last pushed yesterday), licensed MIT. It adds 95 tokens to every session and 5,690 once invoked, about $0.0005 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 2026-09-03.
Other skills, from other repositories
systemic-issue-triage
Trigger: new issue, bug report, triage, backlog, issue flood, community report, root cause, dead-end, blocked user. Attack issues by root class, never one-by-one; fixes must shrink the system, not grow it.
triage-issue
Intelligently triage bug reports and error messages by searching for duplicates in Jira and offering to create new issues or add comments to existing ones. When an agent needs to: (1) Triage a bug report or error message, (2) Check if an issue is a duplicate, (3) Find similar past issues, (4) Create a new bug ticket…
bug-triage
Read all open bugs in production/qa/bugs/, re-evaluate priority vs. severity, assign to sprints, surface systemic trends, and produce a triage report. Run at sprint start or when the bug count grows enough to need re-prioritization.
investigation
Create an investigation ticket + live-doc for sustained root-cause work (retros, bug archaeology, regression hunts).
browse-flows
Browse Power Automate environments and flows interactively. Use when the user wants to browse, list, or explore their flows and environments.
operating-cadence
Designs the rhythm an organization runs on — which reviews happen weekly, monthly and quarterly, what each one decides, who owns the numbers presented, and how a signal at the front line reaches the people who can act on it. Use this to set up a management operating system, fix a meeting calendar that produces no…