long-horizon-prompting
An agent skill by muratcankoylan, from muratcankoylan/Agent-Skills-for-Context-Engineering. Tags: automation, brainstorming, developer-tools, strategy.
What it does
This skill should be used when writing, enhancing, or evaluating the launch prompt for a long-running autonomous agent or a parallel multi-agent orchestration attacking a hard problem: pseudo-formal task briefs that define terms and an exact success predicate linguistically, enumerate non-counting outcomes, set persistence rules with explicit stop and return conditions and effort floors, manage a diverse portfolio of parallel approaches with an approach registry and blocked-route bookkeeping, and gate the return on adversarial audit. Route agent topology and coordination protocols to multi-agent-patterns, runtime control surfaces and loop governance to harness-engineering, evaluator and quality-gate construction to evaluation, judge design to advanced-evaluation, and compaction or memory mechanics to context-compression and memory-systems.
Install
With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:
npx skills add muratcankoylan/Agent-Skills-for-Context-Engineering --skill long-horizon-promptingOr copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):
git clone --depth 1 https://github.com/muratcankoylan/Agent-Skills-for-Context-Engineering
cp -r Agent-Skills-for-Context-Engineering/skills/long-horizon-prompting ~/.claude/skills/long-horizon-promptingSafety box score
Grade B, low risk: 18 of 100 risk points, rated 2026-10-21. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.
This skill is documentation for writing prompts that launch long-running autonomous and multi-agent runs; it contains no executable code or install steps. The main risk is indirect: it encourages strong persistence, anti-stop, and broad orchestration language that can drive expensive autonomous behavior if copied into an active system prompt.
| Category | Level | Why |
|---|---|---|
| Code & shell execution | 0 (none) | No shell commands, scripts, or code evaluation are shipped; the files are Markdown and references only. |
| Network access | 0 (none) | There is no runtime downloading or outbound data flow in the skill itself; the external URLs are citations, not executed fetches. |
| Filesystem reach | 0 (none) | The skill does not read, write, or delete files outside its documentation scope. |
| Secrets & credentials | 0 (none) | It does not request tokens, passwords, keyfiles, `.env` data, or browser sessions. |
| Privilege & persistence | 0 (none) | No sudo, service registration, global installs, or shell/OS setting changes appear anywhere in the shipped files. |
| Instruction hijack surface | 2 (moderate) | The prompt templates include behavior-shaping directives like "Do not tell most agents the currently favored approach" and "Do not return merely because current approaches fail," which are instructions an agent could paste into a live system prompt. |
| Irreversible actions | 2 (moderate) | The skill is explicitly for "a long-running autonomous agent" and "parallel multi-agent orchestration," including "up to 64 concurrent agents" and aggressive persistence, so its purpose inherently expands autonomous reach. |
| Transparency | 0 (none) | All shipped content is readable Markdown with textual references; there are no binaries, minified blobs, or decoded-at-runtime payloads. |
Before installing
- Read `SKILL.md` and `references/task-brief-template.md` to see the persistence and stop/return language you may be adopting.
- If you do not want long autonomous runs, strip or weaken the anti-stop / effort-floor wording before using the skill.
- Check the external URLs in `references/*.md` before relying on them; they are high-volatility citations, not local executable artifacts.
Rated 2026-10-21 by an automated review of 6 files from the skill's repository.
Source
- Repository
- muratcankoylan/Agent-Skills-for-Context-Engineering (all skills from this repository)
- Path
- skills/long-horizon-prompting/SKILL.md
- Branch
- main
- Updated
- 2026-09-19
Related skills
- multi-agent-patterns — This skill should be used when designing multi-agent systems that need context isolation, supervisor or swarm coordination, explicit handoffs.
- goal-prompt — Drafts copy-paste-ready /goal commands for goal mode in Claude Code and Codex.
- bdi-mental-states — This skill should be used when modeling agent mental states with BDI concepts: beliefs, desires, intentions, RDF-to-belief transformations.
- context-compression — This skill should be used when long-running agent sessions need context compression, structured summarization, compaction, token-per-task optimization.
- context-degradation — This skill should be used for diagnosing and mitigating context degradation: lost-in-middle failures, context poisoning, context clash, context confusion.
- context-optimization — This skill should be used for improving context efficiency: context budgeting, observation masking, prefix or KV-cache strategy, partitioning.