Issues

• Add a 'Status' line to the README under the tagline

  enhancementglassbox-agent
  agentic-trust-labs/glassbox-ai#147 · Filter by author sourabharsh opened on Feb 27, 2026
  2

• Cross-agent tracing and per-agent credential isolation

  agentic-trust-labs/glassbox-ai#146 · Filter by author ThinkOffApp opened on Feb 20, 2026
  12

• Bug: max_retries is 0 — agent never retries on failure

  glassbox-agent
  agentic-trust-labs/glassbox-ai#144 · Filter by author sourabharsh opened on Feb 16, 2026
  4

• Bug: fuzzy_find threshold too low (0.3) causes false positive matches

  glassbox-agent
  agentic-trust-labs/glassbox-ai#138 · Filter by author sourabharsh opened on Feb 16, 2026
  5

• Bug: Default reaction is 'confused' instead of '+1' in GitHubClient

  glassbox-agent
  agentic-trust-labs/glassbox-ai#137 · Filter by author sourabharsh opened on Feb 16, 2026
  4

• Bug: HC2 and HC3 descriptions are swapped in constants.py

  glassbox-agent
  agentic-trust-labs/glassbox-ai#136 · Filter by author sourabharsh opened on Feb 16, 2026
  5

• Bug: Off-by-one in CodeEditor.apply() bounds check

  glassbox-agent
  agentic-trust-labs/glassbox-ai#135 · Filter by author sourabharsh opened on Feb 16, 2026
  5

• Bug: temperature_code is set to 1.0 instead of 0.1 in settings.py

  glassbox-agent
  agentic-trust-labs/glassbox-ai#134 · Filter by author sourabharsh opened on Feb 16, 2026
  4

• [Test] Webhook end-to-end verification

  glassbox-agent
  agentic-trust-labs/glassbox-ai#133 · Filter by author sourabharsh opened on Feb 16, 2026
  3

• Update PR body version string from v1 to v2

  glassbox-agent
  agentic-trust-labs/glassbox-ai#131 · Filter by author sourabharsh opened on Feb 15, 2026
  4

• Bug: debate rounds default to 1 instead of 3

  glassbox-agent
  agentic-trust-labs/glassbox-ai#128 · Filter by author sourabharsh opened on Feb 15, 2026
  8

• Bug: trust_floor constant is 0.03 instead of 0.30

  glassbox-agent
  agentic-trust-labs/glassbox-ai#127 · Filter by author sourabharsh opened on Feb 15, 2026
  5

• Bug: get_trust returns 0.50 instead of 0.85 for unknown agents

  glassbox-agent
  agentic-trust-labs/glassbox-ai#126 · Filter by author sourabharsh opened on Feb 15, 2026
  9

• [Fix] Skip redundant syntax check in Tester - tests already catch syntax errors

  glassbox-agent
  agentic-trust-labs/glassbox-ai#114 · Filter by author sourabharsh opened on Feb 15, 2026
  4

• [Fix] Use gpt-4o-mini for Manager classify to reduce latency

  glassbox-agent
  agentic-trust-labs/glassbox-ai#108 · Filter by author sourabharsh opened on Feb 15, 2026
  4

• [Fix] Show before/after diff in JuniorDev comment

  enhancementglassbox-agent
  agentic-trust-labs/glassbox-ai#77 · Filter by author sourabharsh opened on Feb 15, 2026
  8

• [Fix] Add step numbers [1/4] to agent comments

  enhancement
  agentic-trust-labs/glassbox-ai#75 · Filter by author sourabharsh opened on Feb 15, 2026
  5

• [Feature] Extract difficulty from Manager LLM classify response

  enhancementglassbox-agent
  agentic-trust-labs/glassbox-ai#70 · Filter by author sourabharsh opened on Feb 15, 2026
  4

• [Feature] Include test files in agent context for grounded fixes

  enhancementglassbox-agent
  agentic-trust-labs/glassbox-ai#69 · Filter by author sourabharsh opened on Feb 15, 2026
  4

• [Feature] Adaptive EMA alpha — trust learning rate decays with experience

  enhancementglassbox-agent
  agentic-trust-labs/glassbox-ai#66 · Filter by author sourabharsh opened on Feb 15, 2026
  3

• [Bug] get_trust() returns 0.50 for unknown agents instead of 0.85

  glassbox-agent
  agentic-trust-labs/glassbox-ai#49 · Filter by author sourabharsh opened on Feb 15, 2026
  10

• [Eval E10] KeyError: 'critic' missing from EMOJIS dict

  bugeval
  agentic-trust-labs/glassbox-ai#45 · Filter by author sourabharsh opened on Feb 15, 2026

• [Eval E09] Seed data inserts 'critics' instead of 'critic'

  bugeval
  agentic-trust-labs/glassbox-ai#44 · Filter by author sourabharsh opened on Feb 15, 2026

• [Eval E07] ImportError: cannot import '__ver__' from glassbox

  bugeval
  agentic-trust-labs/glassbox-ai#42 · Filter by author sourabharsh opened on Feb 15, 2026

• [Eval E06] ROUND 1 instruction says 'use bullet points' instead of 'no bullet points'

  bugevalglassbox-agent
  agentic-trust-labs/glassbox-ai#41 · Filter by author sourabharsh opened on Feb 15, 2026
  2