Borrowing it
Nothing to install: this file belongs to aget-framework/aget. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/aget-framework/aget/main/.claude/skills/aget-propose-actions/SKILL.mdgit clone --depth 1 https://github.com/aget-framework/agetWrote 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/aget-framework/aget/aget-propose-actions)<a href="https://agentmods.dev/skills/aget-framework/aget/aget-propose-actions"><img src="https://agentmods.dev/badge/skills/aget-framework/aget/aget-propose-actions/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/aget-framework/aget/aget-propose-actions"><img src="https://agentmods.dev/badge/skills/aget-framework/aget/aget-propose-actions.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 9 findings, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 245 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 253 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 259 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 262 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 264 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 289 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 312 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 316 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
- medium Excessive Agency · line 333 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00033 | $0.12257 |
| Opus 5.5 | $0.00013 | $0.04903 |
| Sonnet 5.5 | $0.00007 | $0.02451 |
| Haiku 4.5 | $0.00003 | $0.01226 |
Grade A, and why
aget-propose-actions 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.
Copies of this mod
1 near-identical copy found in the catalogue:
- aget-propose-actions — 89% identical, 0 lines differ
How it starts
The opening of the file, as written. The whole thing — 378 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/aget-propose-actions
Propose ranked next-best actions with evidence grounding, time budgets, and execute-all default. Supersedes the informal "step back. N next-best actions" trigger phrase.
Purpose
The principal's most frequent interaction pattern across the entire fleet (3-5x per session, every agent, every session) is: "step back. N next-best actions, in this session, within the next M mins." Currently each agent interprets this differently — varying in depth, format, evidence citation, and execution behavior. This skill standardizes input parsing, KB review trigger (L335), output format, evidence grounding, parameterization, and the execute-all default (AGENTS.md).
Evidence: 34+ agents use identical phrasing. L637 (Mid-Session Steering via NBAs), L677 (Divergent Proposal Mode), L693 (5-action default), L787 (interpretation variance across agents). #721 (prior issue). 10+ invocations across 6 agents in v3.12.0 lifecycle.
Input
$ARGUMENTS
Parameters (parsed from natural language):
| Parameter | Source | Default | Grammar |
|---|---|---|---|
count |
First number in prompt | 5 (per L693) | "N next-best actions" |
budget |
Time expression in prompt | 30 min | "within the next N mins/hr" |
focus |
Remainder of prompt | Session mandate | "focus on X" / "for Y" |
Authorization-shape flags (mode-gated — REQ-PA-017): the following are documented parameters, but documented ≠ unconditionally accepted. Their acceptance is invoker-mode-gated per Step 2.8: principal-typed → ACCEPT (REQ-PA-014); agent-self-issued → REFUSE (REQ-PA-015, preserves the L976/L979 NBA-fill safety core). Documenting them here promotes them out of "undocumented/aspirational" status (removing the per-session re-litigation friction the F1 arc surfaced) without loosening the agent-mode refusal.
| Flag | Meaning | Default | Acceptance gate |
|---|---|---|---|
count=auto |
Agent picks the count from trigger-evidence (budget = ceiling, trigger-presence = floor) | off (use count numeric) |
principal-mode only |
batch |
Treat the proposed set as one batch-authorizable unit | off | principal-mode only |
go |
Batch-level GO (satisfies Step 4.5); authorizes execution of trigger-evidenced actions only | off | principal-mode only |
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 · +21 lines e82fc230d455
- 1mo ago First seen · 357 lines · 33 tokens per session scan A c59707fa1318
aget-propose-actions is a skill published in the GitHub repository aget-framework/aget (12 stars, last pushed 7d ago), licensed Apache-2.0. It adds 33 tokens to every session and 12,257 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 2026-08-30.
Other skills, from other repositories
aget-close-project
Close a PROJECTPLAN with verifiable-assertion gate (V-tests + commits + retrospective), deferred-surface scan for next-plan handoff, and status transition (ACTIVE → COMPLETE / CLOSED / ABANDONED / SUPERSEDED). Strict counterpart to /aget-create-project (D71 Layer 2). Closes asymmetric verb-pair gap at PROJECTPLAN…
aget-release-critique
Adversarial review with the Critic perspective of the Release Delivery Triad (L818). Dual mandate — spec coverage audit plus deep bug finding, asking what breaks in the field and which assumptions are wrong. Superseded by /aget-check-release.
aget-release-audit-specs
Audit specifications with the Spec Auditor perspective of the Release Delivery Triad (L818) — coverage and consistency mode, asking whether a spec covers all behaviors and where gaps or contradictions remain. Superseded by /aget-check-release.
aget-release-build
Execute release gate deliverables with the Builder perspective of the Release Delivery Triad (L818) — speed-optimized execution asking whether the work is done, fast, and functional. Superseded by /aget-create-release.
aget-file-issue
File issues with L520 governance compliance.
aget-propose-actions
Propose ranked next-best actions with evidence grounding, time budgets, and execute-all default. Formalizes the fleet's highest-frequency interaction pattern.