archive-analysis
Skill304 starsChanged 2 months agoMITView on GitHub
Save completed analyses to the knowledge system's analysis archive for future reference. Use this skill after completing any L3+ analysis, when `/run-pipeline` completes, when the user explicitly says "save this analysis" or "archive this", or automatically at the end of Step 18 (Close the Loop) in the analysis workflow. This skill captures key findings, metrics used, agents invoked, and output file paths so past work can be referenced in future sessions. Trigger whenever you finish validation on a multi-step analysis, complete an analytical deck, wrap up a root cause investigation, finish an opportunity sizing exercise, or close any analysis that produced deliverables worth preserving. Also apply when the user mentions saving work, archiving results, preserving findings, or wants to ensure an analysis can be recalled later. This is your analytical memory system — use it proactively to build institutional knowledge.
In this repository65 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- always-compareSkill
- analysis-designSkill
- analyst-coreSkill
- archaeologySkill
- auth-preflightSkill
- businessSkill
- causalSkill
- chart-to-driveSkill
- close-the-loopSkill
- codex-reviewSkill
- compare-datasetsSkill
- connect-dataSkill
- connect-snowflakeSkill
- context-compareSkill
- context-traceSkill
- data-inspectSkill
- data-mapSkill
- data-profilingSkill
- data-quality-checkSkill
- datasetsSkill
- distribution-profilerSkill
- evalSkill
- evaluate-graderSkill
- experiment-briefSkill
- experimentSkill
- exploreSkill
- exportSkill
- forecastSkill
- google-doc-exportSkill
- google-slides-exportSkill
- guardrailsSkill
- historySkill
- improve-contextSkill
- knowledge-bootstrapSkill
- log-correctionSkill
- maintain-contextSkill
- metric-specSkill
- metricsSkill
- monitor-evalsSkill
- notion-exportSkill
- paceSkill
- patternsSkill
- question-framingSkill
- question-routerSkill
- reliabilitySkill
- resume-pipelineSkill
- run-pipelineSkill
- runsSkill
- score-analysisSkill
- session-handoffSkill
- setup-dev-contextSkill
- setup-notionSkill
- setupSkill
- setup-slackSkill
- setup-snowflakeSkill
- srm-checkSkill
- stakeholder-communicationSkill
- stress-testSkill
- switch-datasetSkill
- The first 60 shown.
The file
Show its rules and text
We split this skill into the rules it gives an agent
30 rules. About 0 are common, 19 are distinctive, and 11 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈19
Distinctive · estimate
This author's own choices. Worth a look.
11
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
After completing an L3+ analysis (post-validation)
- Very specific: line 15
After /run-pipeline completes successfully
- Distinctive: line 24
Check if this is an archive request (create new entry) or a verification request (read existing entry).
- Distinctive: line 29
No new findings mentioned — just wants to see what's already there
- Distinctive: line 54
If validation report exists, extract from there
- Distinctive: line 56
If nothing exists, use findings the user mentioned in their archive request
- Distinctive: line 59
Check validation report or analysis summary for metric references
- Very specific: line 63
Read working/session_state.yaml if pipeline was used
- Distinctive: line 64
Check conversation history for agent invocations (look for "Task" tool uses)
- Very specific: line 67
Run ls -lt outputs/ working/ | head -20 to find recent files
- Very specific: line 71
If no files exist or user didn't mention any outputs, set output_files: []
- Very specific: line 78
Read from working/validation_report.md or working/validation_summary.json
- Very specific: line 79
If validation was not run, set to null and note it
- Very specific: line 81
If actual files don't exist: Create the archive with whatever metadata is available.
- Distinctive: line 130
If file doesn't exist, create it from template:
- Very specific: line 138
Update last_updated to current date (YYYY-MM-DD)
- Very specific: line 139
Write back to index.yaml
- Very specific: line 144
Update last_used to current date
- Distinctive: line 158
When to apply: Only for completed analyses (not partial) with confidence grade B or better.
- Very specific: line 160
After archiving, check if the analysis produced reusable patterns worth saving to .knowledge/query-archaeology/curated/ via …
- Distinctive: line 164
Ask: "Would you like to save any SQL patterns from this analysis?"
- Distinctive: line 166
Only capture queries that passed tie-out or validation checks
- Distinctive: line 177
Ask the user: "Would you like to save any SQL patterns from this analysis?"
- Distinctive: line 178
If the user declines or there are no reusable patterns, skip silently
- Distinctive: line 179
Only offer for analyses with confidence grade B or better
- Distinctive: line 180
Never auto-capture without user confirmation
- Distinctive: line 221
If validation was not run, set confidence to null and note it
- Distinctive: line 230
Analysis index doesn't exist: Create it from template
- Distinctive: line 232
Session state files don't exist: Use conversation history and user-provided metadata to build the archive entry
- Distinctive: line 233
User mentioned files but they don't exist: Only list files that actually exist in outputs/ or working/.
≈ “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
Skill: Archive Analysis
Purpose
Save a completed analysis to the knowledge system's analysis archive for future recall. Captures key findings, metrics used, agents invoked, and output file paths so that past work can be referenced in future sessions.
When to Use
- After completing an L3+ analysis (post-validation)
- After
/run-pipelinecompletes successfully - User says "save this analysis" or "archive this"
- Automatically triggered at the end of Step 18 (Close the Loop)
- VERIFICATION MODE: User says "verify the archive" or "show me what was captured" — read existing entry, don't create new
Instructions
Step 0: Determine Mode
Check if this is an archive request (create new entry) or a verification request (read existing entry).
Verification signals:
- User says "verify", "check", "show me what was captured", "did it save", "was it archived"
- User mentions pipeline "auto-saved" or "already completed"
- No new findings mentioned — just wants to see what's already there
If verification: Skip to Verification Mode (below). Otherwise, proceed to Step 1.
Step 1: Gather Analysis Metadata from Session Context
Record only what exists: findings, metrics, agents, and files come from the files and session
history listed below, or from what the user stated. Anything not found is null or [], never
invented.
Where to look:
- Pipeline state — Read
working/session_state.yamlif it exists (contains pipeline progress, agents run, resume instructions) - Active dataset — Read
.knowledge/active.yamlfor dataset ID - Output files — Scan
outputs/andworking/directories for artifacts created in this session (usels -ltto find recent files) - Validation report — Look for
working/validation_report.mdorworking/validation_summary.jsonfor confidence grade - User's stated findings — If user provides key findings in their request, use those verbatim
What to extract:
- Title: Derive from the original question or business context (check session history for the initial analytical question)
- Question: The original user question (look back in conversation for the first analytical request)
- Question level: From the Question Router classification (L1-L5) — check if router was invoked
- Dataset ID: From
.knowledge/active.yaml - Key findings:
- If validation report exists, extract from there
- Otherwise, check
working/analysis_summary.mdoroutputs/narrative.md - If nothing exists, use findings the user mentioned in their archive request
- Format as single-sentence bullets with numbers
- Metrics used:
- Check validation report or analysis summary for metric references
- Scan SQL queries in
working/*.sqlfor column names - Match against
.knowledge/metrics/index.yamlif available
- Agents used:
- Read
working/session_state.yamlif pipeline was used - Check conversation history for agent invocations (look for "Task" tool uses)
- List in execution order
- Read
- Output files:
- Run
ls -lt outputs/ working/ | head -20to find recent files - Filter for files created during this session (check timestamps)
- Include both deliverables (outputs/) and working files (working/)
- ONLY list files that actually exist — check the directory before listing
- If no files exist or user didn't mention any outputs, set
output_files: []
- Run
- Tags: Auto-generate from:
- Keywords in the original question (mobile, checkout, seasonal, etc.)
- Metric names used
- Dataset name
- Analysis type (funnel, root-cause, segmentation, etc.)
- Confidence:
- Read from
working/validation_report.mdorworking/validation_summary.json - If validation was not run, set to
nulland note it
- Read from
If actual files don't exist: Create the archive with whatever metadata is available. Mark it as partial: true if deliverables are missing.
Step 2: Create Archive Entry
- Read the schema:
.knowledge/analyses/_schema.yamlto understand required/optional fields - Generate a unique ID:
analysis_{YYYYMMDD}_{HHMMSS}(use current timestamp) - Build the entry dict following the schema structure
Example entry format:
- id: analysis_20260403_232623
title: "Mobile checkout conversion drop investigation"
date: "2026-04-03"
dataset: {active_dataset}
question: "What caused the conversion rate drop and how does it relate to the mobile checkout flow?"
question_level: L4
findings:
- "Mobile checkout conversion fell 2.1pp (5.4% → 3.3%) in March; desktop was flat"
- "82% of the drop is concentrated at the payment step on iOS"
- "Sessions with a payment error retry at 11% vs. 64% baseline"
metrics:
- conversion_rate
- checkout_completion
- funnel_drop_off
- device_segmentation
agents:
- question-framing
- data-explorer
- descriptive-analytics
- root-cause-investigator
- validation
- chart-maker
output_files:
- outputs/checkout_conversion_analysis.png
- outputs/conversion_funnel_chart.png
- working/checkout_funnel_analysis.png
tags:
- conversion
- mobile
- checkout
- funnel-analysis
- root-cause
confidence: B
partial: falseStep 3: Append to Index
- Read
.knowledge/analyses/index.yaml - If file doesn't exist, create it from template:
analyses: [] total_analyses: 0 last_updated: null - Append the new entry to the
analyseslist - Increment
total_analyses - Update
last_updatedto current date (YYYY-MM-DD) - Write back to
index.yaml
Step 4: Update Dataset Stats
- Read
.knowledge/datasets/{active}/manifest.yaml - Increment
analysis_count - Update
last_usedto current date - Write back
Step 5: Confirm
Report to user:
Analysis archived: {title}
ID: {id}
Findings: {count} key findings captured
Use `/history` to browse past analyses.Step 6: Capture to Query Archaeology (Optional)
When to apply: Only for completed analyses (not partial) with confidence grade B or better.
After archiving, check if the analysis produced reusable patterns worth saving
to .knowledge/query-archaeology/curated/ via helpers/knowledge/archaeology_helpers.py.
- SQL patterns — If validated SQL queries exist in
working/*.sql:- Ask: "Would you like to save any SQL patterns from this analysis?"
- Offer to capture via
capture_cookbook_entry(title, sql, dataset, tables, tags) - Only capture queries that passed tie-out or validation checks
- Table knowledge — If the analysis revealed useful table metadata:
- Offer to capture/update via
capture_table_cheatsheet(table_name, dataset, grain, primary_key, common_filters, gotchas, common_joins) - Include grain, primary key, common filters, gotchas, and common joins
- Offer to capture/update via
- Join patterns — If the analysis used non-obvious joins:
- Offer to capture via
capture_join_pattern(tables, join_sql, cardinality, validated, dataset) - Record cardinality and whether the join was validated
- Offer to capture via
Rules for this step:
- Ask the user: "Would you like to save any SQL patterns from this analysis?"
- If the user declines or there are no reusable patterns, skip silently
- Only offer for analyses with confidence grade B or better
- Never auto-capture without user confirmation
Verification Mode
When user wants to verify an existing archive (not create a new one):
- Read
.knowledge/analyses/index.yaml - Find the most recent entry (highest index, latest date)
- Display what was captured:
- Analysis ID and title
- Date archived
- Key findings (list them)
- Metrics used
- Agents invoked
- Output files preserved
- Tags
- Confidence grade
- Confirm dataset manifest was updated
- Report archive stats (total analyses, last updated)
Do NOT create a new archive entry in verification mode.
Report format:
Archive verified: {title}
ID: {id}
Status: Successfully archived on {date}
Captured:
- {count} key findings
- {count} metrics tracked
- {count} agents used
- {count} output files preserved
Use `/history` to browse all past analyses.Rules
- Never overwrite an existing archive entry — always append
- Key findings should be one sentence each, factual, with numbers where possible
- Tags should be lowercase, no spaces (use hyphens)
- If validation was not run, set confidence to null and note it
- Archive even partial analyses — mark as
partial: true - Verification requests don't create new entries — read and report existing archives
- DO NOT create standalone analysis markdown files or archive directories — the archive system stores metadata in
index.yamlonly. Output files remain in their original locations (outputs/,working/) and are referenced by path in theoutput_filesarray. DO NOT copy or duplicate artifacts.
Edge Cases
- No outputs exist: Set
output_files: [], do not invent file names. Archive with metadata only. - Pipeline was interrupted: Archive what's available, mark as
partial: true, document reason in areason_incompletefield - Duplicate question: Still archive — different runs may find different things
- Analysis index doesn't exist: Create it from template
- User says 'verify' but no archive exists: Report "No archive found. Would you like to create one?" and proceed to archive mode if confirmed
- Session state files don't exist: Use conversation history and user-provided metadata to build the archive entry
- User mentioned files but they don't exist: Only list files that actually exist in outputs/ or working/. If the user mentioned output files but they're not present, note in confirmation: "Files mentioned but not found: {list}"
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.

