A Claude Code skill that logs rich accomplishments to a daily journal and mirrors summaries to your project's session context.
You ship features, fix bugs, hit milestones — but never write them down. By the end of the week you can't remember what you did. Progress feels invisible.
/accomplish captures what you did with full context: git history, session activity, and your own words. It writes a rich, timestamped entry to your daily log and mirrors a one-liner to your active project context.
/accomplish Shipped the authentication feature to production
That's it. One command. Rich entry with reflection, context, and details — written to ~/Hermes/Daily_Accomplishments/ and mirrored to session-context/.
curl -fsSL https://raw.githubusercontent.com/anombyte93/claude-accomplish-skill/main/install.sh | bashmkdir -p ~/.claude/skills/accomplish
cd ~/.claude/skills/accomplish
curl -fsSL -O https://raw.githubusercontent.com/anombyte93/claude-accomplish-skill/main/SKILL.md
curl -fsSL -O https://raw.githubusercontent.com/anombyte93/claude-accomplish-skill/main/script.py
chmod +x script.pycurl -fsSL https://raw.githubusercontent.com/anombyte93/claude-accomplish-skill/main/install.sh | bashOr check for updates without installing:
bash ~/.claude/skills/accomplish/install.sh --check-update/accomplish <what you accomplished>
/accomplish # prompts you to describe it
/accomplish Shipped the authentication feature to production
/accomplish Just hit $33k AUD revenue from vibe-coding in 3 months
/accomplish Fixed the race condition in the payment queue
- Gathers context — date/time (AEDT), recent git log, file changes
- Builds a rich entry — AI crafts a reflection connecting your accomplishment to your journey
- Writes to daily file —
~/Hermes/Daily_Accomplishments/12-Feb-26.md - Mirrors to session context — one-liner appended to
session-context/CLAUDE-activeContext.md
### 14:30 AEDT
**What**: Shipped authentication feature
**Context**: Key milestone for the Atlas AI platform launch
**Details**: Implemented OAuth2 with JWT, 3 files changed, 2 new tests
**Reflection**: Hard-won after debugging the token refresh loop for 2 hoursThe skill uses a codified architecture — deterministic file I/O is handled by script.py, while AI handles the creative writing (entry composition, reflection, sentiment).
SKILL.md (AI judgment) script.py (deterministic I/O)
├── Ask user what happened ├── gather-context
├── Build rich entry ├── write-entry
└── Craft one-liner summary └── write-mirror
All commands output JSON. Run from any directory.
| Command | Purpose |
|---|---|
gather-context |
Get date/time (AEDT), git log/diff, file paths and existence flags |
write-entry --file PATH --entry TEXT [--create] |
Write/append formatted entry to daily file |
write-mirror --file PATH --line TEXT |
Append one-liner to Accomplishments section in session-context |
| Setting | Default | Description |
|---|---|---|
| Daily directory | ~/Hermes/Daily_Accomplishments/ |
Fixed location for all daily logs |
| Session context | ./session-context/CLAUDE-activeContext.md |
Per-project mirror location |
| Timezone | Australia/Sydney (AEDT) |
Timestamps use AEDT |
- Claude Code CLI
- Python 3.8+
- Git (for context gathering)
MIT — see LICENSE