what-am-i-forgetting
Skill0 starsChanged 7 months agoMITView on GitHub
Comprehensive agenda recall across all of your active projects, phases, gates, queued work, ideas, and cleanups. Casts a wide net across roadmaps, memory files, project TODOs, daily notes, cron manifests, and project folders to produce a consolidated dashboard. Use whenever you ask "what's next", "what's on the agenda", "what am I working on", "what am I forgetting", "what's open", "what's in progress", "give me a recap", or invoke `/what-am-i-forgetting`. Gives a concrete list to disambiguate generic questions about priorities instead of guessing which project is meant.
In this repository24 other files it gives its agents.
- aarSkill
- agent-cost-modelSkill
- banana-makerSkill
- bitter-lesson-scorecardSkill
- boot-tax-monitorSkill
- compensating-complexity-auditorSkill
- context-fork-guideSkill
- context-hygieneSkill
- counterargument-stress-testSkill
- dark-code-auditSkill
- executive-briefingSkill
- failure-asymmetrySkill
- failure-postmortemSkill
- file-intelSkill
- geopolitical-signal-enricherSkill
- get-api-docsSkill
- gh-reviewSkill
- model-routerSkill
- optimize-descriptionSkill
- prompt-rewriterSkill
- skill-auditSkill
- skill-maintenanceSkill
- spec-gap-detectorSkill
- token-burn-auditorSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
14 rules. About 0 are common, 14 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈14
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 22
Use this skill automatically when the user:
- Distinctive: line 30
When the user asks a generic agenda question without naming a project, do not guess which project they mean.
- Distinctive: line 47
Do not omit sections even if empty — write "(none)" so the user can confirm the section was considered, not forgotten.
- Distinctive: line 58
If the user maintains a memory-index file, this is the authoritative source.
- Distinctive: line 68
Look for roadmap/blueprint docs at common locations:
- Distinctive: line 82
Read these to extract pending work items not captured in memory.
- Distinctive: line 89
Look for "Next Actions" sections — these are work items the user wrote down but may not have transferred to memory or open-items yet.
- Distinctive: line 103
Check these locations for running background processes:
- Distinctive: line 118
Every match is a candidate time-boxed item.
- Distinctive: line 159
After producing the full list, add a "Gaps I might have missed" section:
- Distinctive: line 172
Use tables for items with deadlines.
- Distinctive: line 173
Do not narrate — just produce the agenda.
- Distinctive: line 177
If the user corrects an item ("project X is actually done", "you missed project Y", "Z is no longer client work"), update the relevant …
- Distinctive: line 194
If any of these are unset, skip that data source gracefully and flag it in the "Gaps" section rather than failing.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
What Am I Forgetting: Comprehensive Agenda Recall
This skill produces a consolidated view of everything open across your project ecosystem. It is deliberately exhaustive — the cost of missing an item is higher than the cost of listing too many.
When to use this skill
Use this skill automatically when the user:
- Asks "what's next", "what's on the agenda", "what am I working on"
- Asks "what am I forgetting", "what's open", "what's in progress"
- Asks "give me a recap", "where did we leave off", "what's the state"
- Invokes
/what-am-i-forgettingexplicitly - Starts a session with a generic "let's continue" without a specific project
When the user asks a generic agenda question without naming a project, do not guess which project they mean. Produce the full list and let them point at the right item.
Output structure (always use these sections)
- Primary Pipeline / Program Status (if you have a main workstream — phase status + blockers)
- Time-boxed / Expiring Items (table with deadlines, sorted by closest)
- Active Projects — In Flight
- Active Projects — Queued / Planning
- Client / Revenue Work
- Ideas Not Yet Started
- Hygiene / Cleanup Backlog
- Ongoing Automation (crons, daemons)
- Unknown Status (project folders not mentioned elsewhere)
- Suggested Prioritization
Do not omit sections even if empty — write "(none)" so the user can confirm the section was considered, not forgotten.
Data sources to consult (in order)
1. Memory index (always start here, if one exists)
$MEMORY_INDEX (e.g. ~/.claude/.../MEMORY.md, or an equivalent index file)If the user maintains a memory-index file, this is the authoritative source. Skim all sections. Each linked memory file is a potential agenda item. Pay special attention to:
- Sections named "Active Projects" or similar
- Sections for "Paused" / "Early-Stage Ideas"
- Any file whose description contains "expires", "deadline", or a future date
- Any file whose description contains "COMPLETE" / "SHIPPED" (for confirming done)
2. Roadmap files
Look for roadmap/blueprint docs at common locations:
ls $PROJECTS_DIR/*ROADMAP*.md $PROJECTS_DIR/*BLUEPRINT*.md 2>/dev/null
ls $PROJECTS_DIR/*/ROADMAP.md $PROJECTS_DIR/*/BLUEPRINT.md 2>/dev/nullExtract current phase status and open phase items from each.
3. Open-item queues (search broadly)
find $PROJECTS_DIR -maxdepth 3 -type f \
\( -name "NEXT.md" -o -name "TODO.md" -o -name "open_items.json" \
-o -name "open-items.json" -o -name "AGENDA.md" \) 2>/dev/nullRead these to extract pending work items not captured in memory.
4. Recent daily notes (last 7 days)
ls -t $VAULT_DIR/daily/*.md 2>/dev/null | head -10Look for "Next Actions" sections — these are work items the user wrote down but may not have transferred to memory or open-items yet.
5. Project folder inventory
ls -d $PROJECTS_DIR/*/Cross-reference against items already mentioned. Any project folder NOT mentioned in memory, roadmaps, or queues goes into "Unknown Status" — these are often stale folder accumulation that need decisions.
6. Cron / automation manifest
Check these locations for running background processes:
crontab -l 2>/dev/null
ls /etc/cron.d/ 2>/dev/null
ls ~/.claude/triggers/ 2>/dev/null
pm2 list 2>/dev/null
systemctl --user list-units --type=service 2>/dev/null | grep -i runningList what is actively running so the user knows what to monitor.
7. Expiring-item extraction
grep -rE "expires|deadline|by 20[0-9]{2}-" $MEMORY_DIR 2>/dev/nullEvery match is a candidate time-boxed item. Sort by expiration date, soonest first. Flag anything expiring in <=7 days as urgent.
Categorization rules
In Flight = memory/notes say "IN PROGRESS", "ACTIVE", "LIVE", "DEPLOYED", or have a recent daily note with substantive progress.
Queued / Planning = memory/notes say "PLANNING", "SCAFFOLDED", "BLOCKED", or have a NEXT.md but no recent daily note activity.
Client / Revenue = anything tagged with a client name, or living in a
client-engagement folder (e.g. vault/clients/, clients/).
Ideas Not Yet Started = memory description starts with "Early idea", "Flagged, not built", "Theory", "Concept", or similar. No code work done.
Hygiene / Cleanup = stale skill/plugin lists from audits, empty folders, duplicate crons, unreferenced memory files, orphan plugins.
Ongoing Automation = anything in crons/triggers/pm2/systemd that runs unattended. List the name, schedule, and last-known-working date if tracked.
Unknown Status = project folder exists but no memory/roadmap/daily mention. Flag for the user to confirm keep/archive/delete.
Suggested Prioritization rules
Rank by these factors, in order:
- Deadline-driven urgency (any item expiring in <=7 days)
- Blocker on primary pipeline (blocks main-workstream progression)
- Revenue impact (client work, paid engagements)
- Momentum preservation (currently in flight, losing context if paused)
- Hygiene ROI (compound cleanup value)
- Idea maturity (no deadline, lowest priority)
Present the prioritization as a ranked "This week / Next week / Background" triage, not a single long list.
Honest gap-flagging
After producing the full list, add a "Gaps I might have missed" section:
- Projects for which no data was available to assess
- Memory files that couldn't be parsed
- Folders normally skipped (venvs, caches) — confirm they should be skipped
- Sources that should have been checked but couldn't be
This is critical. The skill is only useful if the user trusts it's exhaustive. Explicitly call out what was not checked so they can decide whether to expand the source list.
Output length
Aim for a working dashboard, not a data dump. Roughly 600-1200 words. Use tables for items with deadlines. Use bullet lists for everything else. Do not narrate — just produce the agenda.
Self-update rule
If the user corrects an item ("project X is actually done", "you missed project Y", "Z is no longer client work"), update the relevant memory file immediately and rerun the skill. The skill's value compounds with the quality of the memory index; do not let corrections evaporate.
Configuration
This skill scans user-specific locations. Set these env vars (or inline at invocation):
| Variable | Purpose | Example |
$PROJECTS_DIR | Root of your code projects | ~/projects |
$VAULT_DIR | Your notes vault (daily notes, project files) | ~/vault |
$MEMORY_INDEX | Your memory-index file, if you maintain one | ~/.claude/memory/MEMORY.md |
$MEMORY_DIR | Directory containing memory files | ~/.claude/memory/ |
If any of these are unset, skip that data source gracefully and flag it in the "Gaps" section rather than failing.
Prerequisites (optional but helpful)
This skill is most useful when you already maintain:
- A memory-index file with one-line entries per project/topic
- Per-project
NEXT.mdorTODO.mdoropen_items.jsonfiles - Daily notes in a vault folder
- Roadmap/blueprint documents for major programs
Without these, the skill falls back to project-folder inventory and cron listings, which is still useful for answering "what's running?" but won't capture planned or ideated work.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

