feat: deepen 268 exploitation skills; web session delete; CSS design system; JEV progress checkpoint

agents_md (skills):
- enrich all 255 vulns/ + 13 chains/ agents from thin one-liner stages to
  concrete playbooks: exact tools/commands, per-stack decision points, benign
  proof markers (unique OOB nonces, single reads, URLDNS-before-exec), explicit
  proof criteria, false-positive/pitfall sections, and chaining hooks. Every
  contract preserved (## User/System Prompt, {target}/{recon_json}, FINDING
  block, CWE/Severity, credits). avg 37->53 lines; loader parses all 449.

web console:
- delete a session/report: DELETE /api/runs/:id and DELETE /api/runs (all),
  a Delete button in the run detail and a hover ✕ per sidebar row (tested e2e)
- CSS design system: tokenise the loose values into one scale — 8-step type
  scale (was 10 ad-hoc sizes), radius/z-index/motion/scrim/terminal tokens,
  fix an undefined var(--muted); 66 tokens, 0 loose font sizes, all var() resolve
- stale version labels 4.0.0/4.2.0 -> 4.2.1

harness (JEV / System One):
- typesafe::progress_checkpoint (jev-skill agent-checkpoint pattern:
  continue/pivot/stop) wired into the attack-chain loop to stop looping rounds
  early; works with TypeSafe or local Laya via from_env(); honours --typesafe off
- 390 tests passing

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

agents_md/vulns/oidc_misconfig.md

@@ -8,16 +8,37 @@ You are testing **{target}** for OpenID Connect issuer/nonce/audience validation
 
 **METHODOLOGY:**
 
-### 1. Pull discovery
-- GET `/.well-known/openid-configuration` and jwks_uri
+### 1. Pull discovery + keys
+- `curl -sk https://{target}/.well-known/openid-configuration` → note `issuer`, `jwks_uri`, `id_token_signing_alg_values_supported`, `authorization_endpoint`, `token_endpoint`.
+- Fetch the JWKS: `curl -sk <jwks_uri>` → record `kid`s and key types (RSA/EC).
+- Capture a legitimate `id_token` (decode header+payload with `jwt` CLI or `python -c` base64) to see `iss`, `aud`, `nonce`, `exp`, `alg`, `kid`.
 
-### 2. Test validation
-- Forge id_token with alg=none, wrong iss/aud, reused nonce; swap kid
+### 2. Signature & claim tests (decision points)
+- **alg=none**: strip the signature, set header `{"alg":"none"}`, keep/modify claims → does the RP accept an unsigned token?
+- **alg confusion (RS256→HS256)**: re-sign the token with the PUBLIC key bytes as an HMAC secret → accepted = classic key-confusion.
+- **kid tricks**: point `kid` at an attacker-controlled JWKS (`jku`/`x5u` if honored), path-traversal `kid`, or a `kid` that maps to a predictable key.
+- **iss mismatch**: change `iss` to an attacker IdP → accepted = issuer not validated.
+- **aud mismatch**: set `aud` to another client → accepted = audience not validated (token from a different app replayable).
+- **nonce**: reuse/replay a prior `nonce`, or omit it → accepted = replay possible.
+- **exp**: submit an expired token → accepted = lifetime not enforced.
+- Tools: `jwt_tool <token> -X a` (alg-none), `-X k -pk pubkey.pem` (key confusion), `-I` (claim injection).
 
 ### 3. Confirm
-- Authenticate with a manipulated id_token the RP should reject
+- Proof = a MANIPULATED `id_token` the RP SHOULD reject is accepted for authentication (a session/access token issued, or the authenticated page returned).
+- Show the crafted token header+payload (decoded), the request that submitted it, and the RP's accepting response (mask any resulting session token, keep a prefix).
 
-### 4. Report Format
+### 4. Disprove false positives
+- RP returns `invalid_token`/`signature verification failed`/`invalid_issuer` → validation works → not a finding.
+- alg=none rejected, key-confusion rejected → signature enforced.
+- The token is accepted only by a debug/mock endpoint, not the real RP login → informational.
+- Discovery/JWKS being publicly readable is BY DESIGN → not a finding on its own.
+
+### 5. Chaining hooks
+- Accepted forged token → account-takeover (impersonate any `sub`) and API agents.
+- `aud` confusion → replay tokens between sibling apps sharing an IdP.
+- Attacker-JWKS acceptance (`jku`) → full signing-key control; note the sink.
+
+### 6. Report Format
 For each CONFIRMED finding:
 ```
 FINDING:
@@ -33,4 +54,4 @@ FINDING:
 ```
 
 ## System Prompt
-You are an OIDC specialist. Report only when a manipulated token is actually accepted by the relying party for authentication. Discovery exposure alone is informational.
+You are an OIDC specialist. Report only when a manipulated token is actually accepted by the relying party for authentication — show the decoded crafted token, the submitting request, and the RP's accepting response. If the RP returns invalid_token/signature failure/invalid_issuer, validation works and there is no finding. Discovery or JWKS exposure alone is informational (by design). Keep PoCs benign; do not act beyond proving acceptance.