| name | lit-status |
| description | Session startup status check for academic literature review projects. Reads the project log, counts files in each workflow directory, infers the current phase, and prints a concise status panel — all in one shot. Replaces the manual "check the project before starting" habit that users frequently forget. Use this skill at the start of every session, when resuming after a break, when you're unsure where a project stands, or when a user says "let's continue the review" or "where were we?" without providing context. Works for any lit-review project structure. Invoke with /lit-status.
|
| allowed-tools | Read Glob Grep Bash |
| metadata | {"version":"1.0"} |
Lit-Status — Session Startup Status Check
Run this at the start of a session. The goal is to give the user (and yourself) a
shared, accurate picture of the project state in under 60 seconds, based on what
the files actually say — not what you remember from earlier in the conversation.
This skill is read-only. It does not modify any files.
Step 1 — Locate the Project Root
The project root is the directory that contains the main workflow subdirectories.
Check these in order:
- The current working directory (if it contains LOG.md, CLAUDE.md, or ≥ 3 of the
expected workflow directories)
- Parent directories up to 2 levels up
- If the user named a path in their message, use that
Expected workflow directories (look for any combination of these names):
| Purpose | Common directory names |
|---|
| Search results | search/, results/, data/, raw/ |
| Screening | screening/, screen/, filtered/ |
| Extraction notes | extractions/, notes/, papers/, summaries/, readings/ |
| Synthesis / themes | synthesis/, analysis/, themes/, clusters/ |
| Outline | outline/, outlines/, structure/ |
| Draft sections | draft/, drafts/, sections/, writing/ |
| Final manuscript | manuscript/, output/, final/, ms/ |
| References | references/, bib/, library/ |
If you cannot locate a project root with ≥ 3 recognizable directories, tell the user
and ask them to confirm the working directory before continuing.
Step 2 — Read the Project Log
Look for a session log file: LOG.md, log.md, SESSIONS.md, or journal.md in
the project root.
If found, read the last entry only (the most recent ## YYYY-MM-DD block or
equivalent). Extract:
- Date of last session
- What was done (
Done: line or equivalent)
- What was planned next (
Next: line or equivalent)
If no log file exists: note "No log found — this may be the first session."
Step 3 — Count Files in Each Directory
For each directory found in Step 1, count the relevant files:
ls -1 [dir]/*.md 2>/dev/null | wc -l
ls -1 [dir]/*.csv 2>/dev/null | wc -l
ls -1 [dir]/*.pdf 2>/dev/null | wc -l
ls -1 [dir]/*.bib 2>/dev/null | wc -l
Also check for a manuscript file with quick word count:
wc -w [manuscript_file] 2>/dev/null
For the screening directory specifically, try to extract the number of included papers:
grep -c ",2," [screening_dir]/*.csv 2>/dev/null | tail -1
If that fails, use the extraction note count as a proxy for "papers processed."
Step 4 — Infer the Current Phase
Based on what files exist and how complete they are, assign the most likely phase.
Use this decision logic:
Does a polished/v2/submitted manuscript exist?
YES → Phase 6–7: Polish or submission preparation
Does any manuscript file exist (v1, draft, cited)?
YES and extractions are > 80% of included papers → Phase 5–6: Writing / early polish
YES and extractions < 80% → Unusual — flag as potential gap
Do synthesis files or outline files exist?
YES → Phase 4–5: Synthesis / writing
Do extraction notes exist?
YES, count ≥ 5 → Phase 3: Extraction (in progress or complete)
YES, count < 5 → Phase 2–3: Just started extractions
Does a screening CSV exist?
YES → Phase 2: Screening (in progress or complete)
Do only search result files exist?
YES → Phase 1–2: Search complete, screening not started
No files in any directory → Phase 1: Not started
State the inferred phase as a single line: "Inferred phase: Phase N — [name]"
Step 5 — Print the Status Panel
Use this exact format. Fill in real values; use — for directories that don't exist.
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
LITERATURE REVIEW STATUS · [today's date]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Last session: [date, or "No log found"]
Done: [done line from log, or "—"]
Next: [next line from log, or "—"]
Inferred phase: Phase N — [phase name]
Project files:
search/ [N files] [✓ / ○ / —]
screening/ [N files] [✓ / ○ / —]
extractions/ [N files] [✓ / ● / ○ / —]
synthesis/ [N files] [✓ / ○ / —]
outline/ [N files] [✓ / ○ / —]
draft/ [N files] [✓ / ○ / —]
manuscript/ [N files, word count if exists]
references/ [N .bib files]
Included papers: [N included / N screened, or "unknown"]
Extractions: [N / N_included (X% complete), or just N]
Manuscript: [N words, or "not started"]
Suggested next step:
[one concrete sentence about what to do this session]
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Status symbols:
✓ — directory exists and has files
● — in progress (exists but incomplete relative to expected total)
○ — not yet started (directory exists but empty, or expected but absent)
— — directory not found in this project
Step 6 — Flag Any Blocking Issues
After the panel, add a short "Flags" section if any of these are true:
| Condition | Flag |
|---|
| Extractions < 50% of included papers but draft exists | "Draft started before extraction complete — coverage may be thin" |
| No log file | "No LOG.md found — consider creating one to track session history" |
| Manuscript word count > 15,000 | "Manuscript may exceed journal limits — check target word count" |
[PLACEHOLDER] or [TODO] found in manuscript | "Manuscript contains unfilled placeholders" |
| Extraction count = 0 but screening shows included papers | "Included papers exist but no extraction notes started" |
| Draft directory empty but synthesis directory has files | "Synthesis exists but no draft started — ready to begin writing?" |
Only flag conditions that are actually true. Do not fabricate warnings.
Step 7 — Suggested Next Step
End with one concrete suggested action for today's session, based on the phase and
what looks incomplete. Examples:
- Phase 2: "Screen [N] remaining abstracts in
borderline_review.csv"
- Phase 3: "Extract [N] remaining included papers (currently [X]% complete)"
- Phase 4: "Draft the thematic synthesis outline — extraction is complete"
- Phase 5: "Continue drafting [section name] — [N] sections remain"
- Phase 6: "Run /lit-audit before proceeding to style editing"
- Phase 7: "Replace citation placeholders using the library.bib file"
If the log's "Next:" line is specific and matches the current state, use that instead.
Constraints
- Read-only. Do not write or modify any files during this skill.
- Fast. Prioritize Bash file counts over reading file contents. Only read LOG.md
and a small portion of the manuscript (first 50 words) if needed.
- Honest about uncertainty. If you cannot determine the number of included papers,
say "unknown" rather than guessing. If the phase is ambiguous between two options,
name both: "Phase 3–4 (extraction nearly complete, synthesis not yet started)."
- One status panel, then stop. Do not proactively start the session's work — just
report status and wait for the user to direct the next action.