hermes-insights
starsforks
How to install hermes-insights?
alexai-mcp/hermes-ccc/hermes-insights
$npx -y skills add AlexAI-MCP/hermes-CCC --skill hermes-insightsGenerate insights about your Claude Code usage — what topics you work on most, common patterns, productivity trends.
SKILL.md
| 1 | # hermes-insights |
| 2 | |
| 3 | Analyze your Claude Code usage patterns across sessions and produce a structured insight report: what topics you work on most, which skills you use, what types of projects dominate your workflow, and where your productivity is strongest. |
| 4 | |
| 5 | ## Invocation |
| 6 | |
| 7 | ``` |
| 8 | /hermes-insights |
| 9 | /hermes-insights --days <N> |
| 10 | ``` |
| 11 | |
| 12 | --- |
| 13 | |
| 14 | ## Subcommands |
| 15 | |
| 16 | ### `/hermes-insights` — Full Analysis |
| 17 | |
| 18 | Analyzes all available memory files and session history with no time constraint. |
| 19 | |
| 20 | ### `/hermes-insights --days <N>` — Time-Bounded Analysis |
| 21 | |
| 22 | Restricts analysis to session files created in the last N days. |
| 23 | |
| 24 | **How Claude filters by date:** |
| 25 | |
| 26 | ```bash |
| 27 | find ~/.claude/projects/*/memory/session_*.md \ |
| 28 | -mtime -<N> -type f |
| 29 | ``` |
| 30 | |
| 31 | --- |
| 32 | |
| 33 | ## Data Sources and What Claude Analyzes |
| 34 | |
| 35 | ### 1. Memory Files — Topic and Theme Extraction |
| 36 | |
| 37 | ```bash |
| 38 | ls ~/.claude/projects/*/memory/session_*.md |
| 39 | ``` |
| 40 | |
| 41 | For each session file, Claude reads the `decisions`, `facts_learned`, `artifacts_created`, and `open_issues` sections and extracts noun phrases as topic signals. Topics are then counted and ranked by frequency across all sessions. |
| 42 | |
| 43 | **Output:** Top 10 topics, each with session count and a representative example decision or fact. |
| 44 | |
| 45 | ### 2. Skills Used — Skill Utilization |
| 46 | |
| 47 | ```bash |
| 48 | ls ~/.claude/skills/ |
| 49 | ls ~/.claude/projects/*/skills/ 2>/dev/null |
| 50 | ``` |
| 51 | |
| 52 | Claude counts which skill directories exist and cross-references session memory files for any `/skill-name` invocation patterns mentioned in the `decisions` or `facts_learned` fields. |
| 53 | |
| 54 | **Output:** Skill utilization table (skill name, invocation count estimate, last used date). |
| 55 | |
| 56 | ### 3. Project Types and Domains |
| 57 | |
| 58 | Claude reads each session file's `project:` field (from the hermes-compress YAML) and groups sessions by project slug. It then infers the domain from artifact paths and topic keywords (e.g., `.py` artifacts + "Neo4j" keywords → "graph database / Python backend"). |
| 59 | |
| 60 | **Output:** Project breakdown table (project, session count, primary domain, last active date). |
| 61 | |
| 62 | ### 4. Conversation Patterns |
| 63 | |
| 64 | Claude counts per-session: number of decisions made, artifacts created, problems solved, and open issues left unresolved. These become productivity metrics. |
| 65 | |
| 66 | **Output:** Per-week averages for decisions, artifacts, and resolution rate (problems solved / open issues ratio). |
| 67 | |
| 68 | ### 5. Data Hygiene and Confidence Rules |
| 69 | |
| 70 | - Skip any session file that is missing the expected `hermes-compress` YAML block and record how many files were excluded. |
| 71 | - Normalize topic strings by lowercasing, trimming punctuation, and folding obvious singular/plural variants before counting. |
| 72 | - Prefer explicit evidence from `decisions` and `artifacts_created` over weak inference from prose when assigning project domains. |
| 73 | - Mark a skill as `inferred` if the file only mentions the skill name indirectly and no `/skill-name` invocation is present. |
| 74 | - Downgrade trend claims to `low confidence` when fewer than 5 sessions match the selected date range. |
| 75 | - Report `no sessions matched the filter` instead of fabricating empty charts when `--days <N>` returns zero files. |
| 76 | - Treat duplicate session paths with identical timestamps as one observation so repeated syncs do not inflate counts. |
| 77 | - Fall back to `unknown project` when the `project:` field is absent and artifact paths do not provide a clear slug. |
| 78 | - Separate unresolved carry-over work from newly opened issues so the resolution rate is not overstated. |
| 79 | |
| 80 | --- |
| 81 | |
| 82 | ## Full Output Structure |
| 83 | |
| 84 | ``` |
| 85 | ## Claude Code Usage Insights |
| 86 | Generated: 2026-04-07 | Sessions analyzed: 23 | Date range: 2026-02-14 – 2026-04-07 |
| 87 | |
| 88 | ### Top Topics |
| 89 | | Rank | Topic | Sessions | Example | |
| 90 | |------|--------------------|----------|---------| |
| 91 | | 1 | Neo4j / graph DB | 14 | "Use Neo4j as primary ontology store" | |
| 92 | | 2 | FastAPI / Python | 11 | "REST endpoints for /query and /ingest" | |
| 93 | | 3 | Vercel deployment | 7 | "Deploy to Vercel with vercel link --repo" | |
| 94 | | 4 | OpenCrab ontology | 6 | "Fallback to neo4j when opencrab returns 0" | |
| 95 | | 5 | Discord integration| 4 | "Reply via plugin_discord_discord reply tool" | |
| 96 | |
| 97 | ### Skill Utilization |
| 98 | | Skill | Est. Invocations | Last Used | |
| 99 | |-------------------|-----------------|------------| |
| 100 | | hermes-compress | 18 | 2026-04-07 | |
| 101 | | hermes-memory | 12 | 2026-04-05 | |
| 102 | | hermes-search | 9 | 2026-04-06 | |
| 103 | | hermes-persona | 5 | 2026-04-03 | |
| 104 | | honcho | 3 | 2026-03-28 | |
| 105 | |
| 106 | ### Project Breakdown |
| 107 | | Project | Sessions | Domain | Last Active | |
| 108 | |---------------|----------|---------------------- |
People who liked this also viewed
6 picksType
- find-skillsHelps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skSkillsJul 2026
- grill-meA relentless interview to sharpen a plan or design.SkillsJul 2026
- vercel-react-best-practicesVercel's official collection of agent skillsSkillsJul 2026
- improve-codebase-architectureScan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.SkillsJul 2026
- setup-matt-pocock-skillsConfigure this repo for the engineering skills — set up its issue tracker, triage label vocabulary, and domain doc layout. Run once before first use of the other engineering skills.SkillsJul 2026
- triageMove issues and external PRs through a state machine of triage roles — categorise, verify, grill if needed, and write agent-ready briefs.SkillsJul 2026
