feat: Claude Code command & subagent layer (scope-first workflow)#21
Merged
Conversation
Adds a thin Claude-Code-native front end over the package that packages the authorized workflow (scope -> recon -> validate -> report) with scope enforcement and the scope-compliance audit trail baked in. No new capability; it orchestrates existing modules with authorization and accountability first. - .claude/commands/: /scope (ingest a HackerOne program's scope into a ScopeGuard + bind an AuditLog), /recon (scoped, read-only surface mapping), /validate (minimal-PoC confirmation, cut false positives), /report (remediation-focused writeup with the scope-compliance summary attached). - .claude/agents/: recon-agent, validator, report-writer subagents mirroring the stages, each with a narrow tool scope. - .claude/README.md documents the layer and its guardrails. Every command leads with scope/authorization and routes requests through the audited SecurityScanner, so testing stays in-bounds and recorded. Structure inspired by shuvonsec/claude-bug-bounty (MIT), reduced to the responsible core workflow and wired to this project's ScopeGuard / AuditLog / HackerOneClient. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
A thin, Claude-Code-native front end over the
bugbounty_ctfpackage. It adds no new capability — it packages the authorized workflow (scope → recon → validate → report) as slash commands and subagents, with scope enforcement and the scope-compliance audit trail baked in as the backbone.Slash commands (
.claude/commands/)/scope <program>— ingest a HackerOne program's structured scopes into aScopeGuardand bind anAuditLog. Run first./recon <base-url>— scoped, mostly read-only attack-surface mapping./validate <lead>— confirm a candidate finding with the minimum proof needed; cut false positives./report [path]— remediation-focused writeup with the scope-compliance summary attached.Subagents (
.claude/agents/)recon-agent,validator,report-writer— mirror the stages, each with a narrow tool scope.Why this shape
It ties the whole session's work together and makes it ergonomic:
/scopeuses the HackerOne client (#19) →ScopeGuard;/reconand/validateroute through aSecurityScannerbound to theAuditLog(#17);/validaterecords what worked via the pattern DB (#18);/reportuses the retention-boundedsave_report(#20). The decomposition deliberately reinforces responsible testing — scope first, everything recorded, minimal proof, no escalation, authorized use only — rather than being a broad offensive arsenal.Notes
.claude/— no package/source changes, so it's independent of feat: scope-compliance audit trail for every request #17–feat: bound growing ~/.hermes stores (shared rotation + report retention) #20 and conflict-free. Frontmatter validated.🤖 Generated with Claude Code