feat: implement agent-creator skill and fix multi-tool evaluation

[SyedaQurratAI]

This PR addresses Issue #1120 and includes critical stability fixes:

• Added 'agent-creator' meta-skill for task-specific agent sets.
• Fixed 'evaluation.py' to handle multiple parallel tool calls correctly.
• Added Windows support for 'recalc.py' by using %APPDATA% paths.

Fixes #1120

[SyedaQurratAI]

Hi! I have fixed the path issues for Windows in recalc.py (Issue #1120) and implemented the new agent-creator skill. All recent changes are now verified via web commits. Ready for review! @anthropic-maintainer

[SyedaQurratAI]

Hi @anthropic-maintainer, is there any feedback on this PR? Happy to make changes if needed.

[SyedaQurratAI]

Hi! I have fixed the path issues for Windows in recalc.py (Issue #1120) and implemented the new agent-creator skill. All recent changes are now verified via web commits. Ready for review! @anthropic-maintainer
is there any feedback on this PR? Happy to make changes if needed.

[SyedaQurratAI]

Closing in favor of focused PRs:

• fix: use %APPDATA% for Windows path support in recalc.py #1320 (Windows path fix)
• fifix: handle multiple parallel tool calls in evaluation.py #1321 (evaluation.py fix)
• feat: add agent-creator meta-skill for task-specific agent sets #1322 (agent-creator skill)

[SyedaQurratAI]

Hi, I've split this PR into #1320, #1321, and #1322 as suggested. Could a maintainer take a look when they get a chance?