Shared files describe runtime capabilities and evidence formats. They are
context, not permission to start a runtime, tracker, substrate, network call,
or external mutation.
Just-in-time loading works because a reference read only when a consuming
skill needs it cannot silently become a dependency; anything loaded by default
eventually gets treated as one.
Named failure mode — reference promotion: shared prose quietly outranking
a source skill contract because it was read more recently.
Anti-pattern: citing a shared file as authority for starting a tool or
runtime. Corrective: authority comes from the caller or the consuming skill's
contract; shared files only describe.
Default to the current agent and local shell.
Use a runtime-native fresh context only when the caller or consuming workflow
requests it.
Treat runtime and factory state as adapter evidence; never translate it into
core Plan, Candidate, RPI, or verdict state.
Missing optional tools degrade only the optional capability that needs them.
Source skill contracts and executable behavior outrank shared prose.
The core loop has no hard dependency on this library.
1---
2name: shared
3description: "Shared runtime and evidence references"
4category: ai-agents-and-harness
5source_repo: hashgraph-online/awesome-codex-plugins
6source_path: "plugins/boshu2/agentops/skills-codex/shared/SKILL.md"
7source_url: https://github.com/hashgraph-online/awesome-codex-plugins/blob/HEAD/plugins/boshu2/agentops/skills-codex/shared/SKILL.md
8metadata:
9 source: https://skillmd.com/skills/hashgraph-online/shared
10---
1112# Shared References
1314Shared files describe runtime capabilities and evidence formats. They are
15context, not permission to start a runtime, tracker, substrate, network call,
16or external mutation.
1718Just-in-time loading works because a reference read only when a consuming
19skill needs it cannot silently become a dependency; anything loaded by default
20eventually gets treated as one.
2122Named failure mode — **reference promotion**: shared prose quietly outranking
23a source skill contract because it was read more recently.
2425Anti-pattern: citing a shared file as authority for starting a tool or
26runtime. Corrective: authority comes from the caller or the consuming skill's
27contract; shared files only describe.
2829- Default to the current agent and local shell.
30- Use a runtime-native fresh context only when the caller or consuming workflow
31 requests it.
32- Treat runtime and factory state as adapter evidence; never translate it into
33 core Plan, Candidate, RPI, or verdict state.
34- Missing optional tools degrade only the optional capability that needs them.
35- Source skill contracts and executable behavior outrank shared prose.
3637The core loop has no hard dependency on this library.
3839---
4041**Source:** [`hashgraph-online/awesome-codex-plugins`](https://github.com/hashgraph-online/awesome-codex-plugins) → `plugins/boshu2/agentops/skills-codex/shared/SKILL.md`
Run npx get-skill add hashgraph-online/shared in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.
Shared runtime and evidence references It is listed under Coding & Dev Tools on SkillMD.
This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: makes network calls. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.
This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.
Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.
hashgraph-online (@hashgraph-online) published this skill. Their other Agent Skills are listed on their SkillMD profile.