felipefontoura/axi25-core/axi25-worklog
Claude Code and agent skill by felipefontoura/axi25-core
- community
Checking that the SKILL.md file is still in the repository.
Create and harvest WORK LOGS in AXI25 — dense records of executed technical work (implementation, debugging, deploy, reverse engineering, experiments). Drafted in `00-capture/worklogs/`, then promoted (via harvest) to `35-worklogs/` as strategic raw material mined for content, projects, learning, and connections. Triggers when the user writes or points to a dense work session ("full session of", "~Nh getting X running", "reverse-engineering of", "work log", "logged this session"), or asks to harvest one ("harvest today's log", "what does this log yield", "mine the worklog", "opportunities from this log"). A work log is NOT a journal entry (that is `axi25-journal`, which is LIFE) nor a study note (that is `axi25-study`). NEVER auto-promote a work log to wiki. The harvest is an exploratory intelligence report along a learning journey — NOT fixed knowledge.
Install
npx vskill@latest install felipefontoura/axi25-core/axi25-worklogSKILL.md
Work Log
35-worklogs/ is a first-class layer: the permanent library of dense work sessions, strategic raw material the user mines to extract value — learning, projects, content, and connections — along a learning journey.
Voice: Respond in the language the user is speaking (see 90-system/references/user-profile.md).
Lifecycle (capture → harvest → permanent home)
The raw log is preserved verbatim when it graduates — nothing is lost. Being in 35-worklogs/ means it was harvested; the harvest/ pair is the evidence.
What belongs here (vs neighbors)
35-worklogs/→ WORK I DID. Implementation, debugging, deploy postmortem,
market reverse engineering, experiments. Dense, technical, project-tied.
25-studies/→ STUDY IN FORMATION (axi25-study). Understanding a concept/
book/course. Opens with "studied X", "notes on…", "reflection on…".
40-journal/daily/→ LIFE (axi25-journal). Interior state, mood, family.
Disambiguate by purpose: worklog = OUTPUT (build/execute); study = INPUT (understand). A work log opens with a session ("full session of…", "~Nh getting X running", "reverse-engineering…"). If ambiguous, ASK.
Creating / drafting a Work Log
Drafts live in 00-capture/worklogs/<slug>.md — build incrementally during the session (append as you go). Filename kebab-case in the user's filename_language (default: match the user's language; slug = project/topic); body content in the user's language. NEVER rewrite or summarize the raw content.
Harvesting a Work Log — proactive, confirm first
The harvest turns a raw session into surfaced opportunities along a learning journey, and is the act that PROMOTES the log to 35-worklogs/. The harvest is an exploratory intelligence report ("what did this session yield?"), NOT a to-do inbox and NOT fixed knowledge.
Trigger model: teaser → confirm
When a draft in 00-capture/worklogs/ is finished/mentioned (status: unprocessed), offer a SHORT teaser, then harvest only if the user accepts. Do NOT harvest silently.
*"This log has ~8 opportunities (2 expensive lessons, 3 content seeds, 1 recurringthread with the previous session). Harvest it?"*
If the user explicitly asks ("harvest today's log", "what does it yield"), skip the teaser and harvest directly.
Harvest flow
- Read the work log fully.
- For the Journey lens, scan prior
35-worklogs/*.md(titles + intros, last
~10 by date) to find recurring threads, emerging competencies, and the arc of a problem evolving across sessions.
- Promote the raw log: move it from
00-capture/worklogs/to
35-worklogs/<slug>.md and set status: processed. Preserve the body verbatim.
- Write the report
35-worklogs/harvest/<slug>.md(template below) — same filename. - Append a one-line summary to
log.md.
Do NOT create concept/entity wiki pages from a harvest by default.
Harvest template
The five lenses below are generic. If the user has defined a personal strategy framework in 90-system/references/user-profile.md (e.g. named content channels or business pillars), map the Content/Toolkit lenses onto it; otherwise use the generic labels.
Drop any lens with nothing real to surface — never pad.
Checkbox legend — the checkbox is a conversion tracker (query-able): - [ ] = open opportunity · - [x] = converted (acted + routed) → [[destination]]. Only the 4 actionable lenses use checkboxes; Relations and Journey are analysis (plain - bullets). Concept candidates stay as a slug in backticks (not a ghost wikilink); they become [[link]] only when promoted to the wiki.
The harvest is a REPORT, not an inbox
Each - [ ] is an opportunity. When you ACT on one, it leaves for its real home (content → 30-projects/ or 00-capture/quick/; learning → 25-studies/ or a wiki candidate; kit → a project/toolkit) carrying source: 35-worklogs/<slug>, and gets marked done in the report (- [x] … → [[where]]). The report stays as the permanent record of "raw work → leverage" (e.g. "a 10h session yielded 3 videos + 1 kit"). Update the file status to partial/done.
Routing destinations (for Content & Toolkit)
| Target | Output examples | |
|---|---|---|
| Video / podcast episode | scripted episode from a debugging arc | |
| Thread / post / newsletter | distilled lesson: hook + story + takeaway | |
| Kit (template, script, swipe, agent) | extracted pattern packaged for reuse or sale | |
| Course module | curriculum piece from a real session | |
| Client case study | structured proof of capability | |
| Product feature | technical building block reused | |
| Network asset (intro, pitch) | story snippet from a real session |
Anti-Patterns
- Don't auto-harvest silently — always teaser → confirm (unless directly asked).
- Don't modify the raw log body — the report lives in the
harvest/companion. - Don't promote harvest items to wiki concepts/entities by default.
- Don't treat the harvest as fixed knowledge — exploratory report along a journey.
- Don't let the report rot into an inbox — acted items leave for their real home
AND get marked done; the report is read, not drained.
- Don't run this on a study note (that is
axi25-study). - Don't pad empty lenses to look complete.
Security
- Security scanPassed, 95 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/axi25-worklog/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/felipefontoura/axi25-core/axi25-worklog)