Menu
Review your buffered writing into per-category frequency docs (grammar / rephrasings / idioms / verbs). Usage: /debrief
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Dev: check a target (spec / plan / working diff / code) for architectural drift from the rulebook docs/ARCHITECTURE.md — report qualitative deviations citing rule ids, and triage new ones into docs/Debt.md. Usage: /drift <target>
Push enriched vocab to Anki Desktop as flashcards and pull review progress back. Usage: /anki-sync
Add words/terms you're learning, enriched (translation, alternative translations, examples, synonyms, definition) into your native/learning languages. Comma-separated for several at once. Usage: /loot <word>[, <word2>, ...]
Use whenever you read, query, or mutate shadowling's sqlite database (shadowling.db) — any SELECT/INSERT/UPDATE/DELETE, ad-hoc SQL, dropping/clearing/resetting rows, inspecting data, or checking a migration's result. All DB access goes through a skill entrypoint, a repository method (`models/*`), or `sql.py` — never raw sqlite3. Schema/data-layer changes are NOT made here — they go through an appended migration in appdb.py; this skill only documents that rule and verifies the outcome.
Explain an expression in the language you're learning that you can't read literally — get a verdict (memorize vs learnable rule) + how to read it, saved to the decode dataset. Usage: /aha <phrase> [+ your hunch]
Remove words/terms from your vocab list. Comma-separated for several at once. Usage: /drop <word>[, <word2>, ...]
| name | debrief |
| description | Review your buffered writing into per-category frequency docs (grammar / rephrasings / idioms / verbs). Usage: /debrief |
| allowed-tools | Bash(python3 */debrief.py*) |
Run the deterministic debrief driver as a SINGLE Bash call that begins with
python3 and the full path — the only shape the granted Bash(python3 …)
permission matches (so nothing before it and no chaining):
python3 "${CLAUDE_PLUGIN_ROOT}/debrief.py"
The driver does everything itself — marks drills, triages each session's
languages, runs the five analytical specialists per session, and persists each
session's findings atomically. It streams progress live (flushed) so the run
never goes dark: a marked N drill(s); reviewing N session(s) header, then one
[i/N] <session> … OK / [i/N] <session> … OK (empty) / [i/N] <session> … ERROR <categories> line per session as each completes, then a totals line.
Relay that summary to the user as-is. If any session shows ERROR, tell the user
that a re-run of /debrief retries only the failed sessions. If the driver exits
non-zero with a config notice, point the user at /shadowling:setup. Do not add
analysis or doc contents of your own — the driver is the source of truth.