Bespoke risk assessment for Chocolate Finance. Invoke with /ra to start an interactive assessment, or /ra:publish to push results to Jira.
~/.claude/skills/ra/references/schemas/enums.schema.json existsls ~/.claude/commands/ra/*.mdParse $ARGUMENTS and route:
| Argument | Action |
|---|---|
| (empty) | Invoke /ra:help |
help, --help, -h | Invoke /ra:help |
doctor, --doctor, check | Invoke /ra:doctor |
version, --version, -v | Invoke /ra:version |
update, --update, upgrade | Invoke /ra:update |
assess | Invoke /ra:assess |
publish (with optional flags) | Invoke /ra:publish passing flags |
status | Invoke /ra:status |
| anything else | Invoke /ra:help |
| Environment Variable | Default | Purpose |
|---|---|---|
RA_OUTPUT_DIR | ~/ra-output | Directory for assessment output files |
JIRA_CLOUD_ID | (none) | Required — Atlassian Cloud ID for MCP and API calls |
JIRA_EMAIL | (none) | Required for Jira publication |
JIRA_API_KEY | (none) | Required for Jira publication |
RR_ASSIGNEE_ID | (none) | Optional — Jira account ID for ticket assignee |
| Field | Value |
|---|---|
| Jira Project | RA |
| Cloud ID | $JIRA_CLOUD_ID |
| Issue Types | Epic (Assessment), Task (Finding), Sub-task (Mitigation) |
| High Likelihood | Medium Likelihood | Low Likelihood | |
|---|---|---|---|
| High Impact | Critical | High | Medium |
| Medium Impact | High | Medium | Low |
| Low Impact | Medium | Low | Low |
| Field | Allowed Values |
|---|---|
likelihood | Low, Medium, High |
impact | Low, Medium, High |
risk_rating | Low, Medium, High, Critical |
risk_category | A, B, C, D, ER, F, I, L, O, OO, P, T |
control_type | Preventive, Detective, Corrective |
control_effectiveness | Effective, Partially Effective, Ineffective, Uncertain |
assessment_status | draft, adversarial_reviewed, rectified, discussed, final |
confidence_level | high, medium, low |
epistemic_type | fact, user_claim, assumption, unknown |
subject_type | document_review, planned_change, issue_incident, concept |
source_reliability | high, medium, low, unknown |
Read: ~/.claude/skills/ra/references/workflow/step-1-interview.md
Execute: adaptive conversation to understand the subject, confirm scope, gather materials.
Read: ~/.claude/skills/ra/references/workflow/step-2-ingest.md
Execute: fetch, normalise, record provenance, build subject brief.
Read: ~/.claude/skills/ra/references/workflow/step-3-assess.md
Execute: categorise risks, rate using matrix, epistemic classification, mitigations, projected residual risk.
Read: ~/.claude/skills/ra/references/workflow/step-4-adversarial.md
Execute: challenge against 11 criteria, verify regulatory citations, rectify assessment.
Read: ~/.claude/skills/ra/references/workflow/step-5-discuss.md
Execute: present findings to user, walk through each finding, handle deferred items.
Read: ~/.claude/skills/ra/references/workflow/step-6-output.md
Execute: finalise all files, write assessment_final.json.
All files saved to ${RA_OUTPUT_DIR:-~/ra-output}/<date>_<subject-slug>/:
01_interview.json02_ingest.json03_assessment.json04_discussion.jsonassessment_final.jsonjira_publication.json (created by /ra:publish)