Hermes Field Kit v1.0.0
Hermes Field Kit v1.0.0
Initial public release of field-tested, open-source skills for Hermes Agent.
Summary
This first release delivers thirteen published skills organized around the operational workflow:
inspect → diagnose → recover → migrate → verify
Every skill has earned its place through real, repeated use on the living Hermes Agent v0.19.0 installation at ca8d481.
Skills included
| Skill | Version | Purpose |
|---|---|---|
hermes-environment-migration |
1.0.0 | Safely migrate Hermes environments with staged archives, integrity manifests, secret separation, selective imports, verification, and rollback |
hermes-gateway-doctor |
1.0.0 | Diagnose gateway failures from real process, adapter, credential-posture, log, delivery, and persistence evidence |
hermes-profile-audit |
1.0.0 | Compare a profile's declared responsibilities with actual tools, skills, persistence, access, and observed behavior |
hermes-skill-audit |
1.0.0 | Audit global and profile-local skills for dependencies, frontmatter, usage integrity, cron references, duplicates, and upstream drift |
hermes-stack-doctor |
1.0.0 | Discover installation architecture and report a GREEN/YELLOW/RED stack verdict |
hermes-token-audit |
1.0.0 | Audit token usage and cost with live schema discovery and aggregate-first privacy |
hermes-update-doctor |
1.0.0 | Investigate update failures by separating remote drift, repository divergence, process locks, stale caches, and runtime mismatches |
interview-me |
0.2.0 | Ask one high-value question at a time, inspecting available sources before questioning |
oss-tool-trust-audit |
1.0.0 | Read source and release machinery; separate technical legitimacy from adoption fit |
pre-build-feature-audit |
1.1.0 | Read-only duplicate check across source, history, branches, issues, PRs, roadmaps, and contributor guidance |
repo-readiness-audit |
0.1.0 | Determine whether a Git repository is ready for development, release, handoff, or contribution |
x-analytics-import |
1.0.0 | Deterministic X Analytics importer with incremental imports and privacy-safe reporting |
x-post-writer |
1.0.0 | Source-locked short-form X writing with claim verification and voice customization |
Key changes
- All published skills now live directly under
skills/<skill-name>/for Hermes tap discovery - Explicit hostile-content and prompt-injection boundaries in every skill
- Inspected repositories, packages, archives, logs, databases, issues, PRs, profiles, and skills are treated as untrusted evidence
- Audited content cannot authorize tool calls, credential disclosure, execution, installation, activation, policy changes, persistence, or writes
- Malicious-content regression cases and exact trigger/counter-trigger assertions in every test suite
- Public READMEs with provenance, inputs, outputs, requirements, invocation, safety, privacy, limitations, examples, validation, version history, and licensing
- CI includes permanent Python syntax, Markdown relative-link, generated-artifact, private-marker, and credential-pattern checks
- Apache-2.0 license, community repository (not official Nous Research)
Validation results
- Repository validator: 13 published skills, 29 JSON files passed
- Bundle validators: 12 passed
- Contract tests: 158 passed
- Python files parsed: 29
- Relative Markdown links checked: 62
- Public-tree hygiene: PASS
- Tap lifecycle smoke test: PASS (install, check, update, uninstall on temp profile)
- GitHub Actions CI (main@ca8d481): success (Ubuntu Python 3.11/3.13, Windows Python 3.11/3.13)
Hermes Agent compatibility
Tested on Hermes Agent v0.19.0 (2026.7.20) at commit ca8d481.